Beruflich Dokumente
Kultur Dokumente
Application of biofilm
membrane bioreactor
(BF-MBR) for municipal
wastewater treatment
NTNU
Norwegian University of Science and Technology
Thesis for the degree of Philosophiae Doctor
Faculty of Engineering Science and Technology
Department of Hydraulic and Environmental Engineering
Igor Ivanovic
ISBN 978-82-471-2665-3 (printed ver.)
ISBN 978-82-471-2666-0 (electronic ver.)
ISSN 1503-8181
Doctoral theses at NTNU, 2011:69
Printed by NTNU-trykk
Abstract
The biofilm membrane reactor (BF-MBR) is a wastewater treatment system that
combines biological wastewater treatment with a biofilm process and with membrane
separation for clarification and purification of biologically treated wastewater. Work in
this thesis is experimentally based on laboratory pilot scale systems specially built,
developed and modified for the research purposes in this thesis.
The work in this thesis is presented through six main themes (research chapters (RC)):
RC1: Overview of previous research on knowledge on particle separation in a movingbed-biofilm reactor (MBBR) process
RC2: Overview of previous research on membrane bioreactors (MBR) based on
attached (biofilm) growth.
RC3: Results from empirical studies on the influence of aeration on membrane
performances in the BF-MBR
RC4: Results from mathematical and empirical studies on influence of membrane
reactor design on membrane performances in the BF-MBR
RC5: Results from empirical studies on influence of nitrogen removal on membrane
performances in the BF-MBR
RC6: Results from empirical studies on influence of coagulation and flocculation on
overall performances in the BF-MBR
The main contributions (C) are:
C1:
C2:
C3:
C4:
C5:
Results from the research work in this thesis are based on five main empirical studies
using a pilot scale biofilm MBR (BF-MBR) setup where overall system performance
and membrane performance has been studied. The thesis is structured as a paper
collection based on seven papers, where the first two are review papers and other five
are results of the original research. In addition, there are three publications based on
original research from the study that were published during the early stages of the
project studies, presented only with the references and abstracts as secondary papers.
Abstract in Norwegian
En biofilm membran reaktor (BF-MBR) er et rensesystem for avlpsvann som
kombinerer biologisk rensing ved bruk av en biofilm prosess, og med
membranseparasjon for filtrering og rensing av det biologisk renset avlpsvannet.
Denne avhandlingen er basert p eksperimentelle arbeider med pilot enheter og
laboratoriesystemer som er spesielt bygd, utviklet og modifisert for
forskningsformlene i dette studiet.
Arbeidet i denne avhandlingen er presentert gjennom seks hovedtema (forskning
kapitler (FK)):
FK1: Oversikt over tidligere forskning og kunnskap om partikkelseparasjon i en
moving-bed-biofilm reaktor (MBBR) prosess
FK 2: Oversikt over tidligere forskning p membran bioreaktorer (MBR) basert p
vekst av fastsittende biomasse (biofilm).
FK 3: Resultater fra empiriske studier om hvordan lufting pvirker membranens
prosessegenskaper i en BF-MBR
FK 4: Resultater fra matematiske og empiriske studier om hvordan membranreaktor
design pvirker membranens separasjonsegenskaper i en BF-MBR
FK 5: Resultater fra empiriske studier om hvordan membranprosessen pvirkes av
nitrogenfjerning i en BF-MBR
FK 6: Resultater fra empiriske studier om hvordan koagulering og flokkulering pvirker
den generelle prosessen i en BF-MBR
De viktigste bidragene (B) er:
B1: En oversikt over forskningslitteratur om empiriske studier og kunnskap om MBR
med anvendelse av biofilm prosesser og separasjon teknikker brukt i en movingbed-biofilm reaktor
B2: Foreslag p en metode for definere optimal lufting for membranenheten basert p
minimere mengden av sub-mikron partikler.
B3: Forbedring i membranens ytelsesevne ved endre p membranreaktorens geometri.
B4: Demonstrasjon av mulighetene for en brekraftig drift av foresltte system med
konfigurasjoner av biologisk nitrogenfjerning.
B5: Forbedring av bde membran og generelle prosessytelser ved bruk av forskjellige
tilsetningsstoffer.
Resultater fra forskningsarbeidet i denne avhandlingen er basert p fem empiriske
studier utfrt med en pilot biofilm MBR (BF-MBR) apparatur, der generelle
systemytelse og membranytelse er studert. Avhandlingen er strukturert som en samling
av syv publikasjoner, der de to frste er review artikler og flgende fem artikler er
resultatene fra original forskningen. I tillegg er tre publikasjoner basert p original
forskning fra studiet som ble publisert i en tidlig fase av arbeidet presentert som
sekundre artikler, kun angitt med referanser og sammendrag.
Preface
This thesis is submitted to the Norwegian University of Science and Technology
(NTNU) for partial fulfillment of the requirements for the degree of philosophiae
doctor.
This PhD work has been performed at the Department of Hydraulic and Environmental
Engineering, Faculty of Engineering Science and Technology, Norwegian University of
Science and Technology, with professor TorOve Leiknes as supervisor.
Thesis has been financed by Norwegian Research Council (NRF) / NTNU faculty
fellowship. Experimental part of the thesis was supported by Krger Kaldnes, Norway,
with a biofilm reactor, Zenon GE, Canada with membrane modules and Kemira,
Finland with coagulants.
The thesis is based on experimental research of an advanced system for wastewater
treatment that is comprised of biofilm technology for biological wastewater treatment
and membrane technology for advanced particle separation and purification.
The experimental part of the thesis was carried out in the wastewater treatment
laboratory of the Department of Hydraulic and Environmental Engineering, Faculty of
Engineering Science and Technology, Norwegian University of Science and
Technology.
10
Acknowledgements
First of all I would like to thank to prof. Vojislav Novakovic, NTNU for his support and
encouragement to come to Norway and do PhD studies.
Very special thanks to my advisor Tor Ove Leiknes, NTNU for permanent support from
the very first day of this project, for his patience, encouragement and dedication to this
project.
I would like to thank to my colleagues Jirachote Phataranawik and Cheng Sun, for their
enormous help in wastewater treatment laboratory and for great time we spent in the lab
in last 4 years.
Further more, many thanks to, Thomas Meyn, Astrid Vik Bjrky, Kamal Azrague,
Gema Sakti Raspati, Ciprian Scurtu and Sumihar Silalahi for great, unforgettable time
that we spent together.
Gril Thorvaldsen, Trine Margrete Hrberg Ness, Torgeir Jensen and Arne Grostad,
deserve warmest thanks for helping me in the analytical lab and the workshop.
I would like to thank to my parents and my sister for constant support during very
difficult moments for me, and without their help and support I would have never
finished this work. Thank you.
11
12
Contents
Abstract....................................................................................................................... 5
Abstract in Norwegian............................................................................................... 7
Preface......................................................................................................................... 9
Acknowledgements.................................................................................................... 11
Abbreviations ........................................................................................................... 15
Introduction.............................................................................................................. 17
Problem Outline
Research Context
Research topics (RT)
Overview of papers
Thesis Structure
17
18
18
19
22
23
23
23
37
13
14
Abbreviations
A436
aBF-MBR
BF-MBR
BNR
BOD
CAU
COD
DOC
FCOD
FI
HRT
Ks
LMH
MBBR
MBR
MF
MLSS
MR
MSH-MR
MW
NR
NTNU
NH4-N
NO3-N
OLR
pBF-MBR
PDN
PSD
R
RO
SADm
SBR
SH-MR
sMBR
SMP
SMPc
SMPp
SP
SRT
SS (MLSS)
SUVA
TMP
TN
TOC
15
16
Introduction
By the year 2025 it is expected that 60 % of the world population will live with water
scarcity if current water consumptions remain at the same current level [1]. It is expected
that with further development of human society the need for fresh water will keep
increasing. More than 90% of available fresh water resources will be consumed in the
next fifteen years [2]. Therefore, water reuse and reclamation is inevitable in the years to
come. Current wastewater technologies will have to be upgraded and/or replaced with
new and more advanced water treatment technologies that can provide high quality of
treated water, following sustainable practices, being less expensive to operate while
meeting more strict legal regulations.
The membrane bioreactor (MBR) technology that comprises the activated sludge
process and membrane technology has been recognized as a technology for advanced
wastewater treatment. Several advantages of MBRs over conventional technologies are
recognized, such as the high quality and hygienic effluent achieved, reduced footprint,
lower sludge production, controlled biomass separation and more advanced nutrient
removal ability.
Biofilm process treatment offers several advantages over other wastewater treatment
processes, such as operational simplicity compared to activated sludge, higher biomass
activity due to accumulation of highly specialized microorganisms and higher resistance
of biomass to toxic substances. Implementation of a biofilm process instead of an
activated sludge process leads to creation of a novel treatment configuration, a biofilm
membrane bioreactor (BF-MBR), which has the possibility to combine the best
characteristics of MBR technology and biofilm wastewater treatment processes.
Problem Outline
Implementation of a biofilm reactor instead of an activated sludge reactor in MBR
technology has not been studied in detail, and has been poorly documented in available
literature as of date. Understanding the interaction of the effluent from a biofilm reactor
and the membrane is crucial for sustainable development of this concept. Characteristics
of the effluent from a biofilm reactor differ significantly from that of an activated sludge
process, which ultimately can create different membrane filtration performances.
Therefore, investigation of this interaction through different process configurations and
different operational conditions is important for understanding the potential capabilities
of a BF-MBR process.
___________________
[1] S. Hoffmann, Planet water: Investing in the worlds most valuable resources, 1st Ed. John Wiley and
Sons (2009)
[2] S. J. Khan, S. Hyas, S. Javid, C. Visvanathan, V. Jegatheesan, Performance of suspended and attached
growth MBR systems in treating high strength synthetic wastewater, Bioresource Technology in press
(2010)
17
Research Context
This project has been funded by the Norwegian Research Council (NRF) and partially
funded by AnoxKaldnes. The thesis is a result of experimental work on development of
a new system for advanced wastewater treatment based on coupling a biofilm reactor
and membrane separation technology.
The aim of this study is: 1) to investigate the potential of coupling a biofilm reactor (the
moving-bed-biofilm reactor in particular) with membrane filtration, 2) to give directions
for development of optimal operational conditions that lead to sustainable operations, 3)
to provide a better understanding of membrane fouling phenomena in the proposed
system configuration and 4) to suggest possible methods for membrane fouling
minimization and control.
In order to relate findings from the pilot plants used in this study with real life
situations, it was decided that the complete research work should be done under as real
conditions as possible, i.e. with real municipal wastewater and with commercially
available products such as carriers for biofilm reactor, membrane materials and
commercially available additives.
18
Overview of papers
P1
Particle separation in moving bed biofilm reactor applications and
opportunities,
I. Ivanovic and T.O. Leiknes,
Submitted to journal Separation Science and Technology
Relevance to this thesis: This paper presents a review of findings of studies on
particle separations after a moving-bed-biofilm reactor, which has been used in
experimental research in this study. It describes available conventional and new
techniques for solids/liquid separations after biofilm reactors in wastewater
treatment applications. The paper gives an outlook of application of advanced
techniques such as membrane separations for solid/liquid separations after a
moving-bed-biofilm process
Contribution to the content of the paper: from the review of available
literature, adapted figures (illustrations) based on diagrams in original literature
sources were created, and implementations of new configurations for application
of submerged membrane reactors proposed.
P2
19
P3
Impact of aeration on particle colloidal fraction in the biofilm
membrane reactor (BF-MBR)
I. Ivanovic and T.O. Leiknes
Published in Desalination 231 (2008) 182-190
Relevance to this thesis: This paper presents findings of experimental studies
on BF-MBRs within research topic one (RT01). Submicron particles are
identified in early studies as a main contribution to membrane fouling. This
paper investigates the impact of aeration intensity on the characteristics of the
submicron particles resulting in the membrane reactor and suggests a simple
method for estimating sufficient membrane aeration for sustainable operation
with minimizing enlargement of submicron particles.
Contribution to the content of the paper: I designed and performed
experimental part, performed data analysis and wrote the most of the paper.
P4
Membrane reactor as a tool for better membrane performance in a
biofilm - MBR (BF-MBR)
I. Ivanovic and T.O. Leiknes
Published in Desalination and Water Treatment 25 (2011) 259267.
Relevance to this thesis: This paper presents findings of experimental studies
on BF-MBRs within research topic two (RT02). Reduction of membrane fouling
by alternative reactor designs was studied where mathematical calculations were
experimentally verified. This paper suggests alternative reactor designs to a
completely mixed flow reactor typically used in MBRs as a tool for enhanced
membrane filtration performance.
Contribution to the content of the paper: I designed and performed
experimental part, performed data analysis and wrote the most of the paper.
P5
Impact
of
denitrification
a biofilm-MBR (BF-MBR).
on
the
performance
of
20
21
Thesis Structure
The thesis is structured in six chapters and two appendixes as follow:
Chapter 1: Introduction
Chapter 2: State of the Art
Chapter 3: Material and methods
Chapter 4: Results
Chapter 5: Evaluation of Results
Chapter 6: Conclusion and future work
Appendix A: Selected papers
Appendix B: Secondary papers
22
Paper P2:
The biofilm membrane reactors (BF-MBR) - a reviewWith development of MBR technology, asides from main stream efforts, several
research groups have been working to supplement or replace biological suspended
growth in MBR with attached growth, searching for new possibilities to overcome
limitations of conventional MBR processes based on activated sludge. The total number
of internationally available publications in the field of biofilm applications in MBR
shows that this topic has not yet received significant attention in the research
community in comparison with the amount of research work related to conventional
MBR. The second paper (P2), gives a review and systematization of research work done
on MBR with biofilm implementations. A classification of MBR systems with biofilm
implementations in two generic groups is proposed:
1. assisted biofilm MBR (aBF-MBR) - where addition of attached (biofilm)
growth to conventional activated sludge
MBR (AS-MBR) assist
biodegradation, and
2. pure biofilm MBR (pBF-MBR) where a biofilm reactor exclusively
performs the biodegradation without presence of activated sludge
Implementations of biofilms in MBR have been shown as mostly successful in terms of
enhanced nutrient removal and mainly improved membrane performances. However
some studies showed a negative impact of applied biofilm, which was elaborated and
commented in the paper.
23
24
Figure 1. Basic configuration of BF-MBR used in the thesis for experimental research
Table 1. Basic characteristics pilot plant that was subject in this study
Moving-bed-biofilm reactor
Number of reactors (in series): 4
Volume of each reactor: 65 L
Type of carrier: K1: Filling fraction 67 % of reactor volume
Surface area for biofilm growth - 335 m2/m3,Total growth area - 84.5 m2
Membrane reactor
Zenon ZW10 Membrane Pilot Module
Type of membrane Hollow fiber with configuration outside/in
Nominal pore size 0.04 micron
Volume of membrane reactors 9-41 L
Operating characteristics
Flux 22 - 52, Backwash flux 38-58 LHM
Operating cycle: 4.75 min. production and 0.25 min. backwashing
Recovery ~ 90-96 %
25
26
Results
Summary of the research papers
Research paper P3:
Impact of aeration rates on particle colloidal fraction in the biofilm membrane
bioreactor (BF-MBR)
In this study a method for estimation of aeration rate in BF-MBR is proposed, taking in
account permeability drop for a certain aeration rate and distribution of particles
affected by the intensity of aeration. Different aeration rates were tested (from highest to
lowest) with a daily shift towards lower aeration rates. The measured change in TMP
(i.e. permeability decline) was related to the amount of the highest fraction of
submicron particles measured in the reactor. Published literature results suggest that
there is a threshold rate of air scouring intensity above which no significant effect or
improvement of membrane performance is observed. However, operating the membrane
with air scour rates close to the threshold value could have a negative effect on particle
size distribution, inducing higher amounts of submicron particles, which have been seen
as potential foulants responsible for irreversible membrane fouling. Therefore, a simple
and quick to perform method was proposed where peak values from particle size
distribution (PSD) measurements and measured fouling rates were plotted against
specific aeration demand to identify which could be the most suitable operating
condition. This approach represents an alternative strategy to defining optimal
operational conditions for the system with respect to aeration rates. In this study, for the
system configuration that was subject of the study, optimal aeration rates were found to
be not lower then 1.68 and not higher then 3.37 Nm3m-2h-1.
Figure 2 illustrates the procedure for defining an optimal specific aeration demand
based on measurements and specifications for the pilot plant used in this study.
10
0,20
0,18
0,14
0,12
9
0,10
0,08
0,06
PSD [number %]
0,16
0,04
0,02
8
0,00
0
5
3
SADm [Nm /m h]
Figure 2. Fouling rate and PSD values of the most present submicron fraction for
different aeration rates (SADm)
27
0,25
0,20
0,15
SH-MR
0,10
MSH-MR
0,05
0,00
0
10
12
14
16
18
20
22
24
26
28
30
32
Time [hour]
Figure 3. TMP before and after backwash for CM-MR, SH-MR, MSH-MR.
Production flux 35 LMH, and backwash flux 38 LMH
28
Pressure (bar)
0.1
0
-0.1
PDN config.
NR config.
-0.2
-0.3
After backwash
-0.4
Before backwash
-0.5
0
10
15
Time (days)
20
25
30
29
0.14
Fouling rate [bar/day]
0.12
)
2
SMPp: R = 0.6155
0.1
DOC: R2 = 0.5518
0.08
0.06
0.04
FCOD: R = 0.7217
0.02
0
0
10
20
30
40
50
Amount retain by membrane [mg/L]
60
30
2,5
2
1,5
1
y = 0,58x + 49,30
R2 = 0,97
200
150
Ref
100
Al 9ppm
0,5
0
0,01
y = 1,50x + 52,91
R2 = 0,97
250
TMP [mbar]
y = 3,97x + 36,83
R2 = 0,99
300
Inlet
BF eff.
Al 9ppm
Fe 22,5ppm
3,5
50
0,1
10
100
Fe 22.5ppm
Li
1000
Size [m]
20
40
60
80
100
120
140
(R f)
160
31
180
32
Evaluation of results
Findings from the research topic 1 (RT01) suggest a new method for estimation of
optimal aeration rates with respect to minimizing the amount of submicron particles in
the membrane reactor, which are considered detrimental for membrane filtration in a
BF-MBR process. Method is simple to perform and gives a relatively straight forward
result that can be used in optimizing the intensity of air scouring as an important fouling
control method. However, the method requires further refinements where on-line
monitoring of particle size changes as a function of changes in aeration rates could give
more accurate predictions and thus optimal operation. The investigation did not cover
the relation between aeration intensity and biochemical parameters responsible for
membrane fouling such as soluble microbial product, total organic carbon or chemical
oxygen demand, so supplemental studies are required to be done.
In the research topic 2 (RT02) a new approach in modifying the membrane reactor
design that lead to improved membrane performance was proposed. By integrating a
flocculation zone and sludge hopper under the membrane module it is possible to reduce
the amount of suspended and soluble organic matter, and to some extent reduce the
amount of submicron particles, around the membrane. This led to lower fouling rates
and improved overall membrane filtration performance. Validation of the proposed
concept should be done on larger or full scale.
Results from research topic 3 (RT03) showed that nitrogen removal in a BF-MBR
process can be easily and efficiently done by implementation of a post-denitrification
mode of operation. Stable denitrification rates with average nitrogen removal of about
80% were done by keeping the C/N ratio at 4.5. In comparison to a nitrification scheme,
implementation of post-denitrification caused higher membrane fouling potential where
pore narrowing and blocking was suspected to be a dominant mechanism. Results point
out that residual organic matter from a post-denitrification configuration has higher
concentration of low MW organic matter which was considered as a main reason for the
higher fouling rates measured. Further investigations of the nature of organic residuals
after biological treatment are required for a better understanding of the fouling nature
and dominant fouling mechanisms in this configuration.
In the research topic 4 (RT04), an evaluation of five different commercial coagulants
was performed in order to investigate the indirect effect these may have on membrane
fouling. All tested additives showed an ability to coagulate/flocculate colloidal organic
matter and furthermore reduce membrane fouling. Polymerized alum had at optimum
dosages the highest ability to destabilize colloids and reduce colloidal organic matter.
Iron chloride showed a better ability to flocculate suspended matter around the
membrane area. The enlargement of floc sizes and floc surfaces in the upper-micron
range caused a reduction of submicron particles by adsorption on/in the flocs formed.
Filterability and dewaterability of the suspension with added iron chloride was better
than with the other four additives tested. Iron chloride showed the best performances
and therefore is suggest as additive of choice. This study showed a good potential of
modified cationic polymers as fouling reducer in BF-MBR.
33
34
Conclusion
This study demonstrated the feasibility of implementing two independently developed
technologies, MBBR and MBR, into a novel configuration called a biofilm-MBR (BFMBR). Sustainable membrane operations over a period of one month with two different
biological configurations at relatively high production fluxes (i.e. 30 LMH) and high
recovery rates (i.e. >95%) without frequent chemical cleaning were demonstrated in
paper 5 (P5). Effluent from a post-denitrification scheme showed a higher fouling
potential due to a higher low MW organic content. A cyclic mode of operation with a
short periodical backwash, with backwash flux around ~130% or higher of the
production flux, and moderate air scouring rates was observed to give good membrane
fouling control and sustainably long operation cycles of the PDN configuration.
Control of membrane fouling by reduction of particulate matter, both suspended and
colloidal with new membrane reactor designs was demonstrated in paper 4 (P4). The
proposed designs investigated in this study included introducing an integrated
flocculation zone in the membrane reactor coupled with a sedimentation zone beneath
the membrane module. The modified membrane rector design provided a significantly
lower concentration of MLSS and COD around the membranes, and subsequently a
more sustainable membrane performance due to much lower overall fouling rates.
Optimization of aeration for membrane scouring with respect to particles size analysis is
demonstrated in paper 3 (P3). The results show the importance of finding the
relationship between sufficient aeration to minimize membrane fouling, while
preventing formation of colloidal particles due to excessive shear forces caused by the
aeration. An approach to defining optimal operating conditions with respect to aeration
rates is proposed. By evaluating overall fouling rates and PSD analysis based on an
assessment of the differential number percent of submicron particles, recommendations
for a design aeration rate may be defined. Optimal aeration rates found for the pilot
plant configuration used in this study was found to be in the range of 1.68 to 3.37 Nm3
m-2 h-1.
Addition of different additives and their effect on reduction of colloidal and soluble
matter was investigated in paper 6 (P6). Their effect on particle agglomeration and
reduction of submicron particles was further investigated in paper 7 (P7). The best
improvement in membrane performance was observed for the higher chosen dosage of
iron chloride, while polymerized alum gave lower improvement at the same applied
dosage. This was related to the highest reduction in surface area of submicron particles
in the range 0.04 0.3m and lower organic colloidal content. A larger surface area and
larger flocs, probably formed a protective cake layer on the membrane surface which
has significantly improved membrane performances at higher iron dosage. Based on the
research conducted in this study the iron chloride coagulant in particular appears to be
the additive of choice for a BF-MBR process.
35
36
Future Work
Results from this study demonstrate that it is possible to maintain a sustainable
operation of a biofilmMBR concept as a result of coupling of two independently
developed technologies, MBBR and MBR, at various operational conditions.
Future work should include experimental evaluation of two main topics:
1) measurements for improved membrane operations, and
2) measurements for improved biological treatment
As this study has been conducted on a laboratory scale pilot plant system, all findings
and results should be verified on full scale pilot plant systems.
1)
There is a need to experimentally investigate the effect of the production cycle length on
minimizing of reversible and consequently irreversible fouling, since all experiments in
this study were done with a production cycle of 4.75 minutes and 0.25 minutes
backwashing. Investigation of optimal cycle length is expected to lead to improved
membrane performances and lower operational cost.
Furthermore, there is a need for optimization of the backwash intensity and length of a
backwash cycle. This would lead to better control of irreversible fouling and better
overall performance.
The effect of aeration intensity as a tool for fouling control requires more detailed
investigations since in this study it was demonstrated that there is a potential hazard for
membrane fouling when improper empirically based aeration rates are employed.
Additionally, the effect of intermittent aeration should also be studied for this
application since all experiments in this study are conducted with continuous aeration
rates. Future work should include knowledge from conventional MBR studies where the
length of intermittent aeration may be optimized.
Backpulsing, i.e. very short (several seconds) with high intensity backwashing, can
potentially be a good fouling control strategy as it has the capability to reduce fouling
caused by pore blocking and pore narrowing of submicron particles. This technique has
not yet been tested in a BF-MBR configuration and is an interesting option that should
be investigated.
The effect of membrane relaxation as a strategy for fouling mitigation was not included
in this study. Studies with AS-MBR have shown that this is an interesting alternative,
however, its applicability to the BF-MBR process needs to be assessed. Furthermore,
the investigations in this research work were based on continuous extraction of excess
sludge. A combination of intermittent sludge extraction with membrane relaxation is
seen as having a good potential for more advance membrane operation which could lead
to lower membrane fouling rates
37
Membrane cleaning is a topic that has received relatively little attention in MBR
technology in general. Investigating the possibility for alternative membrane cleaning
strategies that can avoid classical chemicals for membrane cleaning (e.g. sodium
hypochlorite) is recommended. Evaluation of strategies for zero discharge of chemicals
used for membrane cleaning purposes from BF-MBR is therefore a topic of future work.
Investigation of an alternative process configuration with cascade series of submerged
membrane reactors with completed mixed flow as proposed in Paper 1 (P1) is
recommended. The proposed concept has design simplicity, could be easy to operate,
would provide high effluent quality at each stage and has flexibility regarding system
operations with frequent peak loads.
2)
It was demonstrated that lower organic loading rates (OLR) on the biofilm reactor
creates a more favorable effluent characteristic which has good filterability and induces
lower fouling rates. A more accurate assessment of the relationship between OLR and
membrane fouling should be studied.
Optimization of consumption of aeration for biological treatment is an issue for future
work. Combination of fine bubble aeration with mechanical agitation of suspended
media could lead to lower energy consumption for the oxidation process in the biofilm
reactor.
Removal of nutrients (i.e. nitrogen and phosphorus) has been studied in a post
denitrification configuration and by applying chemical precipitation. The evaluation of
the effect of pre-denitrification and optionally a combination of pre- and postdenitrification should be studied in the future.
Biological phosphorus removal in continuous flow schemes with implementation of
both biofilm and activated sludge is also a topic that warrants future investigations in a
BF-MBR system.
Application of different biofilm support media that can offer the possibility of
simultaneous nitrification and denitrification should be tested. Furthermore, there is a
media that offers good phosphorus removal without need for recirculation loops which
could be an interesting alternative to media used in this study.
Application of a membrane aerated biofilm reactor for biological treatment offers a
good possibility for reduction of air consumption needed for mixing suspended carriers,
and also offers the possibility for simultaneous nitrification/denitrification in deeper
layers of biofilm formed on membrane support layer. Applying alternative biofilm
reactors such as this is an interesting development of the BF-MBR concept that warrants
future investigations.
38
A possible BF-MBR process configuration with pre coagulation and rapid biological
secondary treatment suitable for diluted wastewater (e.g. Norwegian conditions) is
another interesting alternative that should be evaluated.
39
40
41
42
43
58
Paper P2: The biofilm membrane bioreactor (BF-MBR) - a reviewI. Ivanovic and T.O. Leiknes
Submitted to journal of Desalination and Water Treatment
59
60
Abstract
Membrane bioreactors (MBR) based on the activated sludge process is a relatively new
technology, with implementation worldwide increasing over the last 20 years. In
parallel to commercial development, a lot of research work has been done in
fundamental studies, development and optimization of this technology. Although the
main focus has been on activated sludge processes, several research groups have been
investigating biofilm based MBR systems. The biofilm processes have several
advantages over activated sludge process and can be used as complementary, assisted to
activated sludge MBR (aBF-MBR) or self standing, pure biofilm based MBRs (pBFMBR). This article reviews the status of MBR technology with biofilm
implementations. Reports published within the last 10 years are reviewed with respect
to aBF-MBR and pBF-MBR studies, highlighting advantages proposed of this approach
over activated sludge MBRs, identifying performance and operational characteristics
given, and taking an outlook of perspectives in further development of this concept.
Keywords: biofilm membrane bioreactor, biofilm carriers, membrane fouling
1. Introduction
The MBR technology, for both municipal and industrial wastewater treatment, has seen
significant growth in the last 10 years, boosted by a need for more advanced wastewater
treatment, more strict legalization and increasing scarcity of fresh water resources [1].
This technology is primarily based on the conventional activated sludge concept where
secondary clarifiers are normally replaced by submerged low pressure polymeric
membranes for solid/liquid separation. The MBR technology has gained popularity due
to several outstanding advantages over conventional process i.e. high quality effluent
(very often hygienically highly purified), lower footprint, lower net sludge production,
and improved nutrient removal [2]. Major disadvantages of the process are membrane
fouling - which limits sustainability and wider applications, higher energy demand mostly caused by air scouring demand, and higher capital costs due to the price of
membranes. Better understanding of membrane fouling mechanisms, optimization of
energy consumption and cheaper membrane materials have overcome some of these
disadvantages, making this technology even a more realistic and viable choice by the
end of this decade [3][4][5][6].
61
62
opportunity to reduce SRTs for full nitrification. Shorter SRT require higher biofilm
surface, where typical SRT values for an IFAS process are between 2 and 6 days.
Subsequently, addition of carriers with biofilm growth reduces the SRT of an activated
sludge bioreactor, which leads to increased capacity of the bioreactor or opens the
possibility to operate the bioreactor on lower mixed liquor suspended solids (MLSS).
This approach is beneficial for upgrading activated sludge processes. From Figure 1, it
can further be seen that addition of a high biofilm surface area will lead to sufficient
biodegradation where activated sludge is no longer needed. This is typically the case in
a moving bed biofilm reactor (MBBR) where degradation becomes a function of the
biomass available on the carriers and HRT, virtually without suspended matter (i.e. no
activated sludge). The high biofilm surface area in MBBR is typically achieved by
adding biofilm carries with a high surface area at high volumetric filling fractions,
typically up to 2/3 of the reactor volume [13].
500000
Lower temperatures and
Lower COD/TN ratio
400000
300000
200000
100000
MBBR
MCRT=HRT=3-20h
IFAS
MCRT=2-6 days
ACT.SLUDGE
MCRT > 6 days
0
0
Figure 1. Threshold of required biofilm surface area for reaching full nitrification for
typical municipal wastewater as a function of MCRT (SRT) adapted from [7]
By applying the model illustrated in Figure 1 it can be concluded that by adding biofilm
surface area it is possible to decrease the SRT of an activated sludge reactor,
consequently increasing the capacity of existing reactors. It is also possible to design the
treatment plant based on a pure biofilm process without applying activated sludge (i.e.
decoupling SRT and HRT) if a large enough surface for biofilm growth is provided.
There are different kinds of media (i.e. biofilm carriers) that can be used for the
applications described above. Examples of types of commercially available media
typically used in IFAS and MBBR applications, with basics characteristics, are given in
Table 1.
63
Table 1. Several commercial types of media mostly used in full scale plants worldwide,
adapted from [8]
Name of media
Type of bed
Ringlace
BioWeb
RBC
Linpor Sponge
Captor Sponge
Kaldnes K1
Fixed
Fixed
Moving
Moving
Moving
Moving
Typical filling
fraction [%]
25-50
25-50
n/a
15
5-15
15-70
64
Improvment [%]
15
10
5
0
n/a
n/a
-5
-10
-15
-20
COD NH4 TN
TP
COD NH4 TN
TP
5,1
13
20
COD NH4 TN
TP
COD NH4 TN
65
TP
66
67
68
particles are caught by larger particles that were retained by the membrane separation
and further settled in a sedimentation zone at the bottom of the membrane reactor.
Application of pBF-MBR for shipboard wastewater treatment (including oily bilge
wastewater) was demonstrated in a study by Sun et al. [42][43][44][45]. Good and
stable biodegradability of oil and other organic compounds was ensured by application
of this process using both very compact dead-end and side stream schemes. A great
recovery capacity of the pBF-MBR process from oil and salt shock loads in the feed
water was demonstrated. The prefered process configuration was found to be a sidestream design that employs membranes with tighter pores and combined sedimentation
beneath the membrane unit.
12
10
8
6
Higher OLR
Higher
aeration rates
Inlet wastewater
Biofilm effluent
Membrane reactor
Lower OLR
Higher HRT
Lower OLR
Lower
aeration rates
4
2
0
0,01
0,1
10
69
alternative hydrodynamic arrangement the pBF-MBR system was able to produce low
MLSS (~ 100 mg/L) in the reactor, opening the possibility to use UV inactivation as a
fouling control method. Addition of a UV unit resulted in 24% lower fouling rates.
An alternative approach has been proposed where a compact tertiary membrane
treatment as a polishing step after a moving bed biofilm reactor in combination with
disc filtration and flotation (DAF) [49]. This approach relies on a sequential removal of
detached biomass from biofilm reactor, first disk filtration and/or flotation unit,
resulting in low solids loads on the membrane unit. Given approach resulted in
membrane fluxes in the range of 40-80 LMH being achieved. However, this approach
adds another unit of operation in the treatment train and higher configuration
complexity. In addition, the reported higher cleaning frequency and use of coagulant
and cationic polymer for the membrane filtration are obvious drawbacks of this
approach.
A recent comparative study with a fixed bed pBF-MBR and AS-MBR was conducted
by Ng et al., where a 71 % lower production of total SMP (60% less carbohydrate and
77,6 % less total protein) in the pBF-MBR compared to the AS-MBR was
demonstrated, which resulted in 25-30% higher fluxes for the biofilm process [50]. This
study further demonstrates the potential of a biofilm process compared to an activated
sludge process applied to membrane bioreactor technology.
Conclusions
Implementation of a biofilm process for wastewater treatment is beneficial due to the
potential of simplicity for operation compared to activated sludge, higher biomass
activity, higher resistance of the biomass to toxic substances / shock loads, and
development of a higher biodiversity of the microorganisms responsible for the
biological treatment. Although not commonly commercially available to date, biofilm
processes in membrane bioreactor technology have been shown to be potentially
beneficial. Application of biofilm processes in AS-MBRs is beneficial due to the ability
of the biofilm process to reduce the high SRT and MLSS values typically required for
complete biodegradation of constituents in the wastewater. Inclusion of a biofilm
process is practically achieved by addition of a support media that provides a high
surface area for biofilm growth. Higher specific surfaces area and higher filling
fractions are desirable since this can lead to more compact bioreactors and increase
capacities of existing activated sludge systems. The aBF-MBR can achieve the same
organic removal and nitrification rates as comparable AS-MBR designs. Higher total
nitrogen and total phosphorus removals can also be achieved within a single throughput process. Other benefits include the potential for simultaneous nitrification and
denitrification (i.e., through existence of anoxic/anaerobic zones in the deeper layers of
the biofilm component), smaller floc sizes, higher microbial activity and more diverse
microbial community present in biofilm which mainly contribute to improved nutrient
removals. Filterability is generally reported as improved, where lower fouling rates and
higher fluxes were observed in most studies. Less bound EPS and less SMP are
generated or adsorbed by the biofilm, which is considered to lead to improved
membrane performances. Reduction of MLSS concentration is certainly a desirable
70
71
[12] W. N. Lee, I. J. Kang, C. H. Lee, Factors affecting characteristics in membranecoupled moving bed biofilm reactor, Water Research, 40 (2006) 1827 - 1835
[13] H. degaard,. A new moving bed biofilm reactor - applications and results. Water
Science and Technology 29 (1994) (10-11): 157165
[14] X. C. Wang, Q. Liu., Y. J. Liu, Membrane fouling control of hybrid membrane
bioreactor: Effect of extracellular polymeric substances, Separation science and
technology, 45, (2010) 928-934
[15] Q. Liu, Z. C. Wang, Y. Liu, H. Yuan, Y. Du, Performance of a hybrid membrane
bioreactor in municipal wastewater treatment, Desalination 258 (2010) 143 147
[16] S. J. Khan, S. Hyas, S. Javid, C. Visvanathan, V. Jegatheesan, Performance of
suspended and attached growth MBR systems in treating high strength synthetic
wastewater, Bioresource Technology in press (2010)
[17] K. Sombatsompop, C. Visvanathan, R.B. Aim, Evaluation of biofouling
phenomenon in suspended and attached growth membrane bioreactor systems,
Desalination 2010 (2006) 138-149
[18] Q.Y.Yang, T.Yang, H.J.Wang, K.Q.Liu, Filtration characteristics of activated
sludge in hybrid membrane bioreactor with porous suspended carriers,
Desalination 249 (2009) 507-514
[19] Z. Liang, A. Das, D. Beerman, Z. Hu, Biomass characteristics of two types of
submerged membrane bioreactor for nitrogen removal, Water Research 44 (2010)
3319-3320
[20] S. Yang, F. Yang, Z. Fu, R. Lei, Comparison between a moving bed membrane
reactor and conventional membrane reactor on organic carbon and nitrogen
removal, Bioresource Technology 100 (2009a) 2369-2374
[21] S. Yang, F. Yang, Z. Fu, R. Lei, Comparison between a moving bed membrane
reactor or and conventional membrane reactor on membrane fouling, Bioresource
Technology 100 (2009b) 6655-6657
[22] X. Huang, C. H: Wei, K. C. Yu, Mechanism of membrane fouling control by
suspended carriers in a submerged membrane bioreactor , Journal of membrane
science 209, (2008) 7-16
[23] C. H. Wei, X. Huang, C. W. Wang, X. H. Wen, Effect of suspended carrier on
membrane fouling in a submerged membrane bioreactor, Water Science and
Technology, vol 53 No6 (2006) 211 220
[24] H.H. Ngo, M.C. Nguyen, N.G. Sangvikar, T. T. L. Hoang, S. W. Guo, Simple
approach towards a design of an attached-growth sponge bioreactor (AGSB) for
wastewater treatment and reuse, Water science and technology 54 (2006) 191-197
[25] C. Psoch and S. Schiewer, Direct filtration of natural and simulated river water
with air sparginig and sponge ball application for fouling control, Desalination
197 (2006), 190-204
[26] W. Yang, N. Cicek, J. Ilg, State-of-the-art of membrane bioreactors worldwide
research and commetcial applications in North America, Journal of membrane
science 207 (2006) 201-211
[27] A. Drews, M. Vock, U. Bracklow, V. Iversen, M. Kraume, Does fouling in MBRs
depend on SMP? Desalination 231 (2008) 1-3, 141-149
[28] X.Y.Li, S.F.Yang, Influence of loosely bound extracellular polymeric supstances
(EPS) on flocculation, sedimentation and dewaterability of activated sludge,
Water Research 41 (2007) 1022- 1030
72
73
74
75
76
Abstract
In BF-MBR as in all sMBR systems sufficient aeration is necessary to maintain local
cross-flow conditions for sustainable membrane operation. Aeration rate is normally
based on previous experiences and manufactures recommendations. BF-MBR with an
external submerged membrane unit, designed only for particle separation, could be
relatively easily optimized in terms of membrane aeration. Combining observed fouling
rates with measured particle size distribution and minimizing both parameters indicates
a desirable sustainable aeration rate for given wastewater characteristic, operating
parameters (i.e. production/backwash flux, recovery) and membrane module geometry.
Keywords
Biofilm-MBR, colloidal fraction, fouling rate, particle size distribution (PSD), specific
aeration demand membrane area (SADm)
INTRODUCTION
Biofilm membrane bioreactor (BF-MBR) is a hybrid system for municipal wastewater
treatment, where the moving-bed-biofilm reactor (MBBR) for biodegradation of soluble
organic matter is coupled with a submerged membrane reactor (sMBR). BF-MBR has
the potential to utilize the best characteristics of a biofilm process and membrane
separation resulting in compact, efficient particle removal schemes. As in all membrane
processes, fouling is a main challenge that needs to be addressed. The efficiency of the
process is constrained by the accumulation of materials on the surface of, or within the
membrane resulting in a reduction in membrane permeability/production and the overall
performance of the treatment process.
Aeration in MBR systems (both for biological and membrane operational purposes) is
the most cost demanding factor in terms of energy consumption. Air scouring is
necessary in submerged membrane systems in order to generate localized cross-flow
conditions along the membrane surface thereby reducing cake deposit on the membrane.
However, the relationship between aeration and flux or transmembrane pressure decline
is still not fully understood. Aeration rates in MBR systems are based on previous
experience and normally recommended by membrane suppliers (Judd, 2006).
77
78
unit
mg/L
mg/L
Avg.
114.45
274.5
142.2
22.4
Max
286.11
382
180
48.1
Min
45.79
184
80.2
9.13
g O2/m2d
4.53
6.32
3.04
g O2/m2d
2.35
2.98
1.32
mg/L
156.75
204
30.5
438.33
224
45.2
<1
62.22
140.1
27.2
mg O2/L
mg O2/L
mg/L
The membrane reactor was designed as an external submerged membrane reactor with a
volume of 33 L. A ZW-10 pilot plant membrane module supplied by Zenon
Environmental Inc. was applied in this study. The membrane unit was operated under a
constant flux of 52 L.m-2.h-1 and it was periodically backwashed with 5.8 % of the
79
permeate production. A constant backwash flux of 58 L.m-2.h-1 was applied. During the
operating period wastewater temperature was between 8.2 and 15.7 oC. HRT of the
membrane reactor was less than 45 min. From the bottom of the membrane reactor
retentate was extracted continuously with a flow rate ~2L/h. Recovery in the membrane
reactor was set to 95.6 %. The filtration was operated in a cyclic mode (5 minutes)
consisting of 4.75 minutes production followed by 0.25 minutes backwash. Continuous
coarse bubble aeration was applied to promote local cross-flow conditions along the
membrane surface. The aeration system is integrated into the membrane module at the
bottom of the membrane bundle, designed as a four small holes with diameter of 2 mm.
Membrane module specifications and operating modes are summarized in Table 2.
Table 2. Membrane module specifications and operating mode
Membrane reactor
Zenon ZW10 Membrane Pilot Module
- Type of membrane hollow fiber
- Nominal pore size 0.04 m
- Configuration outside/in
Volume of membrane reactor 33 L
Operational characteristics
Flux 52 L.m-2.h-1,
Backwash flux 58 L.m-2.h-1
Total permeate production 43.2 L/h
Concentrate flow - 2 L/h
Operating cycle: 4.75 min. production
0.25 min. backwashing
Recovery = 95.6 %
Aeration rates: 0.84, 1.68, 3.37, 5.05, 6.74
Nm3m-2h-1
Analysis
Particle size distribution (PSD) analysis of the wastewater was done by using laser
diffraction spectroscopy (Beckman Coulter LS230). The development of
transmembrane pressure (TMP) was measured continuously using an online pressure
transducer connected to a data acquisition system from National Instruments, FieldPoint
(FP1000 and FP-AI-110), with the LabVIEW data acquisition and analysis software.
TMP and temperature was logged for every two second. Data series where then
extracted from the raw data depending on analysis, e.g. for overall performance the
values from the beginning of a production cycle were extracted and plotted over time.
Suspended solids (SS) were analyzed by filtering through a Whatman GF/C 1.2 m
according to the NS 4733. Chemical Oxygen Demand (COD) and ammonia (NH4-N)
were measured with the Dr Lange LCK 114, 314, 302 and 303 cuvette tests. For the
Filtered Chemical Oxygen Demand (FCOD) samples were first filtered with Whatman
GF/C glass microfiber filters (1.2 m). Sludge volume index (SVI), capillary suction
time (CST) and time-to-filtrate (TTF) were preformed according to the Standard
methods in order to evaluate settling, dewatering and filtering characteristics of the
retentate. The performance of the membrane filtration unit in the BF-MBR has been
investigated with five different aeration rates; 0.84, 1.68, 3.37, 5.05 and 6.74 Nm3/m2h
expressed as specific aeration demand membrane area (SADm) according to Judd
(2006). The aeration is expressed as SADm, however, as the pilot plant is based on a
small-scale unit the values of the SADm measured do not represent values one can
80
expect from full-scale systems and the trends and relationships of the values are
therefore of more essence in this study.
RESULTS AND DISCUSION
Different aeration rates did not affected the characteristics of permeate due to the fact
that stable permeate quality was observed during the whole experimental period.
Furthermore it was not possible to observe any effect of different aeration rates on other
parameters of treated wastewater e.g., SS/VSS, CST, TTF, SVI etc. The measured
quality parameters for the retentate and permeate from the membrane reactor are
summarized in Table 3.
Table 3. Characteristic of retentate and permeate from membrane reactor
Retentate
Permeate
parameter
SS
VSS
COD
CST
TTF
SVI
SS
FCOD
NH4-N
Turbidity
unit
mg/L
mgO2/L
(second)
(ml/g)
mg/L
mgO2/L
mg O2/L
NTU
Avg.
1211,39
289,07
1209,11
43,18
56
124,53
22,6
Max
2583,74
636,40
2057,1
166,70
180
215,03
0
30,5
<1
< 0,1
Min
568,79
127,67
511,2
15,60
44
47,00
17,2
81
280
0,84 Nm3/(m2h)
1,68 Nm3/(m2h)
3,37 Nm3/(m2h)
5,05 Nm3/(m2h)
6,74 Nm3/(m2h)
260
Permeability [LMH@20oC/bar]
240
220
200
180
160
140
120
100
15
30
45
60
75
90 105 120 135 150 165 180 195 210 225 240 255 270 285 300
Time [second]
82
300
Permeability [LMH@20oC/bar]
250
200
150
100
50
3
6,74 Nm /(m h)
5,05 Nm /(m h)
3,37 Nm /(m h)
1,68 Nm /(m h)
0,84 Nm /(m h)
0
0
Time [days]
83
10
9
8
7
6
6,74 Nm3/(m2h)
5,05 Nm3/(m2h)
3,37 Nm3/(m2h)
1,68 Nm3/(m2h)
0,84 Nm3/(m2h)
5
4
3
2
1
0
0,01
0,1
Figure 4. PSD as a function of aeration rates zoom at submicron area below 0.1 m
One approach to defining the optimal mode of operation with respect to aeration rates is
to evaluate the relationship between membrane fouling rates and the effects of aeration
on suspended solids characteristics. Results from PSD analyses combined with
measured fouling rates for different SADm values could potentially indicate the
desirable range for recommended aeration rates for the most efficient mode of operation
for a system. The recommended SADm would ultimately define the condition that
provides the least energy consumption for aeration while maintaining minimum fouling
rates. Based on the results from this study, overall fouling rates measured as a function
of SADm have been plotted together with PSD analysis results for the same conditions
(Figure 5). From the presented results in Figure 5 a desirable range for SADm should be
estimated for values higher than 1.68 but lower than 3.37 Nm3.m-2.h-1, for the given
membrane reactor design and chosen operating conditions.
It is important to note that these values are specific for the system and operating
conditions applied in this study. However, the approach represents and alternative
strategy to define optimal operating conditions for a system with respect to aeration
rates. In combination with the flux-stepping approach to define the optimal flux range
of operation, analysis of PSD effects and fouling rates as a function of SADm may be
applied to optimize membrane aeration rates. Studies are currently being conducted to
generalize this approach and validate this analysis as a means to define recommended
operating conditions with respect to aeration rates for a given system.
84
0,20
10
0,18
0,14
0,12
0,10
0,08
0,06
PSD [number %]
0,16
0,04
0,02
0,00
8
0
5
3
SADm [Nm /m h]
Figure 5. Fouling rate and PSD values of the most present submicron fraction for
different aeration rates (SADm)
CONCLUSIONS
The aim of this study was to define a procedure that can determine sustainable aeration
rates for a BF-MBR process with a given membrane reactor design, sustainable flux and
wastewater characteristics. The results show the importance of finding the relationship
between sufficient aeration to minimize membrane fouling, while preventing formation
of colloidal particles due to excessive share forces caused by the aeration. An approach
to define optimal operating conditions with respect to aeration rates is proposed. By
evaluating overall fouling rates and PSD analysis based on an assessment of the
differential number percent of submicron particle, recommendations for a design
aeration rate may be defined. Further studies are currently being conducted to develop
this approach and validate a general form and applicability as a design tool for optimal
and sustainable operation of the membrane filtration unit for a BF-MBR process.
ACKNOWLEDGMENT
AnoxKaldnes, Norway, for support with the biofilm reactor and ZENON Environmental
Inc., Canada, for supplying the membrane modules.
85
REFERENCES
Bick A., Yang F., Shandalov S. and Oron G., (2007). Data envelopment analysis for
assessing optimal operation of an immersed membrane bioreactor equipped with a
draft tube for domestic wastewater reclamation, Desalination, 204, 17-23
Bouhabila, E.H.; R. Ben Am and H. Buisson (2001). Fouling characterisation in
membrane bioreactors, Separation and Purification Technology 22, 123-132
Defrance, L., M.Y. Jaffrin, B. Gupta, P. Paullier and V. Geaugey (2000) Contribution of
various constituents of activated sludge to membrane bioreactor fouling,
Bioresource Technology, 73, 105-112
Ji L., Zhou J., (2006) Influence of aeration on microbial polymers and membrane
fouling in submerged membrane bioreactors, Journal of Membrane Science, 276,
168-177
Juang L.C, Tseng D.H.,Yin Lin Y.H. (2007). Membrane processes for water reuse from
the effluent of industrial park wastewater treatment plant: a study on flux and
fouling of membrane, Desalination, 202, 302-309
Judd S., The MBR Book Principles and Applications of membrane bioreactors in
water and wastewater treatment.(2006) Elsevier Ltd.
Le-Clech P., Jefferson B., Judd S.J. (2003). Impact of aeration, solids concentration and
membrane characteristics on the hydraulic performance of a membrane bioreactor,
Journal of Membrane Science, 218, 117129
Leiknes T., degaard H. (2007). The development of a biofilm membrane bioreactor
Desalination, 202,135-143
Leiknes, T.; Bolt H., Engmann M., degaard H. (2006). Assessment of membrane
reactor design in the performance of a hybrid biofilm membrane bioreactor (BFMBR). Desalination; 199, 328-330
Liu, R., Huang, X.,Sun, Y.F. and Qian,Y. (2003) Hydrodynamic effect on sludge
accumulation over membrane surfaces in a submerged membrane bioreactor.
Process Biochemistry, 39(2): p. 157-163.
Melin, E.; Leiknes, T., Helness H., Rasmussen V., degaard, H. (2005). Effect of
organic loading rate on a wastewater treatment process combining moving bed
biofilm and membrane reactors. Water Science and Technology; 51(6-7):421-430
Park, J.-S., K.-M. Yeon, and C.-H. Lee, (2005) Hydrodynamics and microbial
physiology affecting performance of a new MBR, membrane-coupled highperformance compact reactor. Desalination, 172(2): p. 181-188.
Psoch, C., S. Schiewer, (2006) Anti-fouling application of air sparging and backflushing
for MBR. Journal of Membrane Science, 283(1-2): p. 273-280.
Standard Methods for the Examination of Water and Wastewater (1998). 20th edn,
American Public Health Association/American Water Works Association/Water
Environment Federation, Washington DC, USA.
Ueda, T., Hata,K., Kikuoka, Y., Senio, O. (1997) Effects of aeration on suction pressure
in a submerged membrane bioreactor. Water Research, 31(3): p. 489-494.
Wisniewski G. ,Grasmick A. (1996). Floc size distribution in a membrane bioreactor
and consequences for membrane fouling, Colloids and Surfaces A:
Physicochemical and Engineering Aspects, 138, 403-411
86
87
88
89
90
Abstract
Coupling biofilm reactors with membrane filtration as biofilm membrane bioreactors
(BF-MBR) is an interesting alternative technology to activated sludge membrane
bioreactors (AS-MBR). Biofilm technology for wastewater treatment can provide a
substantially lower suspended solids environment for membrane filtration compared to
activated sludge processes. Potential benefits are; less membrane clogging / sludging
problems, lower fouling potentials, ease of membrane cleaning, reduced energy
consumption for air-scouring, and new membrane module/reactor designs. This study
was aimed to investigate alternative membrane reactor designs as a tool to improve
membrane performance in a BF-MBR process. Three different designs were
investigated. A simplified model was developed to predict and analyze the performance
of the membrane reactor designs chosen. Results showed that solids control can be
achieved, in particular the MLSS concentration, as well as a reduction of the colloidal
submicron particle fraction, thereby reducing membrane fouling. Modification of the
membrane reactor in a BF-MBR process is beneficial. The alternative designs
investigated in this study included introducing an integrated flocculation zone in the
membrane reactor coupled with a sedimentation zone beneath the membrane module.
The modified membrane rector design provided a significantly lower concentration of
MLSS and COD around the membranes, and subsequently a more sustainable
membrane performance due to much lower overall fouling rates.
Keywords: biofilm membrane bioreactor, membrane reactor, suspended solids control,
membrane fouling
Introduction
A Biofilm MBR (BF-MBR), based on coupling a biofilm reactor (BF) and a submerged
membrane reactor (sMR) is an alternative concept to conventional MBR systems based
on an activated sludge process for advanced wastewater treatment. The concept of the
BF-MBR process has previously been investigated by combining a moving bed biofilm
reactor (MBBR) followed by a submerge membrane reactor (MBR) [1][2]. An
operational challenge of submerged AS-MBR systems is that the process deals with
liquors having high concentrations of total solids as well as dissolved compounds such
as extracellular polymeric substances (EPS) leading to membrane fouling. Air scouring
is commonly applied to prevent clogging and fouling of the membrane modules, an
energy intensive component of AS-MBR systems. Possible advantages of the BF-MBR
91
concept lie in the fact that biomass is attached to suspended carriers and there is no need
for sludge (i.e. biomass) recirculation in the system. The amount of surplus biomass that
become detached from the biofilm carriers also generate much lower suspended solids
concentrations to be separated in membrane reactor[1][2][3][4]. Subsequently, the lower
amount of suspended matter that needs to be separated gives lower viscosity, less
fouling potential, i.e. cake deposition and clogging and less biofouling, thus reduction
of the energy consumption required from the air-scouring system needed for fouling
control and mitigation [3][4]. This characteristic opens the opportunity for designing a
submerged membrane reactor that can operate with relatively low concentrations of
suspended matter (MLSS) thereby overcoming some of the key bottlenecks in AS-MBR
processes. Several references in the literature can be found where the effect of MLSS
concentration on membrane performance (i.e. membrane fouling or permeability
decline) has been evaluated [4][5][6]. It was shown that in AS-MBR higher MLSS
concentrations induce higher viscosity and higher fouling potential [8], increased cake
layer fouling and resistance [9], and decreased normalized permeability [6]. In general
these studies report better membrane performances (i.e. less fouling) when the
concentration of MLSS is lower [10][11]. However, not only the concentration of the
suspended material is of significance but also the composition and characteristics of the
material, in particular the colloidal fraction [11][12][13][14] .The effect of bio-solids
concentration (i.e. MLSS) in the BF-MBR has not been fully investigated, though
previous studies have shown a correlation between lower fouling rates when lower
MLSS concentrations are observed around the membranes, where the significance of the
colloidal fraction was demonstrated [15][16][17][18]. The potential benefit of the BFMBR process combined with the low solids load to the membrane reactor is the
opportunity to design and operate the membrane unit for enhanced particle removal and
thus improved membrane fouling mitigation and control.
The membrane reactor design combined with mode of operation is an important aspect
with respect to the characteristics and MLSS concentrations that can be achieved around
the membranes. In this study the impact of alternative membrane reactor designs and
operating modes has been investigated as tool for improving the membrane filtration
performance in a BF-MBR. The approach has been to design the membrane reactor for
improved solids control to reduce fouling and to investigate how this may affect the
characteristics of the solids around the membrane, in particular the colloidal fraction.
Three different membrane reactor designs have been investigated and operated at pilot
scale under varying operating conditions.
Theory
In submerged systems the membrane modules are designed either as externally
submerged units or directly immersed in the bioreactor [5]. In both cases the principle
reactor design and configuration will be the same, as illustrated in Figure 1. A
conventional approach of defining flow and mass balances can be applied to describe
and analyze the reactor. In submerged membrane reactors operation commonly includes
air-scouring for fouling control and recycling of the biomass between the
biodegradation stage and the membrane filtration stage, both resulting in the reactor
configuration functioning as a completely mixed membrane reactor, CM-MR.
92
Q out1; c1
Qin ; c 0
V ,c
Q out 2 ; c 2
dc
=Qin c0 -(Qout1 c1 +Qout2 c 2 )
dt
(1)
Where
V - volume (m3),
c - concentration of MLSS (kgm-3),
Qin - flow rate in (m3d-1),
Qout - flow rate out (m3d-1),
0 - subscript: inlet
1 - subscript: permeate,
2 - subscript: concentrate,
Assuming no suspended matter in the permeate, i.e. c1=0 and completely mixed
conditions, i.e. c2=c , the concentration of MLSS inside the membrane reactor can be
expressed as:
c(t)=co
Qin
Q
+co (1- in ) e
Qout2
Qout2
Qout2
t
V
(2)
or,
c(t)=
- 1 t
co
1 - R e SRT
1-R
(3)
93
V
Qout2
(4)
and R is recovery,
Q
R=1- out2
Qin
(5)
SRT=
By analyzing the reactor design and operating conditions using equation (3) it is
possible to predict the expected performance as a function of flow rates (i.e. recovery)
and suspended solids load. For each operating condition one can then determine the
steady state concentration of MLSS in the membrane reactor based on MLSS
concentration and recovery. The results of the analysis for a given condition are shown
in Figure 2.
10000
MLSS [mg/L]
R=99%
R=98%
R=97%
R=95%
1000
R=90%
R=80%
R=70%
R=50%
R=10%
100
0
20
40
60
80
100
Time [hour]
94
adding a factor (Ks) that takes into account the reduction of MLSS due to the modified
reactor geometry.
An expected concentration of MLSS around the membrane can be calculated based on
equation (1) including Ks and expressed as:
c(t)=co
Qin
Qin
+co (1)e
K SQout2
K SQout2
KSQout2
t
V
(6)
MLSS [mg/L]
Ks=1
1000
Ks=4
Ks=10
100
0
20
40
60
80
100
Time [hour]
95
Methods
Three membrane reactor designs were chosen and compared in this study: 1) a
conventional completely mixed reactor (CM-MR), 2) a membrane reactor with
integrated sludge hopper (SH-MR) and 3) a membrane reactor with a modified sludge
hopper design (MSH-MR). Illustrations of the three reactor properties and differences
are shown in Figure 4. The study was conducted with small-scale pilot plant setups
using Zenon ZW 10 pilot plant membrane modules, and with the three reactor volumes
of 9, 27 and 41 L respectively. Each membrane unit was feed with effluent from a pilot
plant MBBR consisting of four moving-bed-biofilm (MBBR) reactors installed in
series. The volume of each reactor was 65 L and each reactor was filled with biofilm
carriers type K1, with specific surface area 335 m2/m3, supplied by Krger Kaldnes.
Filling fraction was 67 % of reactor volume with total surface area for biofilm growth of
84.5 m2. Details and specifics of the pilot plant configuration and setup have been
reported in previous studies [16][17].
The BF-MBR pilot plant configuration was operated first with a low strength municipal
wastewater, and later with a high strength wastewater which was mixture of municipal
wastewater and synthetic wastewater (Table 1).
During operation with the low strength wastewater, the membranes were operated in a
cyclic mode consisting of a 4.75 minutes production time and a 0.25 minutes backwash
cycle. Production flux was set at 35 L.m-2.h-1 and backwash flux at 38 L.m-2.h-1, recovery
96% and specific aeration demand of SADm ~ 3.37 Nm3m-2h-1. The MBBR reactor was
operated with a 4 hour HRT, giving on average the quality parameter values and
treatment efficiencies as shown in Table 1.
a) CM-MR
b) SH-MR
c) MSH-MR
Figure 4. Illustration of the three alternative membrane
reactor designs investigated in this study with Zenon
ZW-10 membrane module
96
During operation with the high strength wastewater, the membranes were operated with
a constant flux of 22 L.m-2.h-1 (no backwash or relaxation), recovery 96%, and the
specific membrane aeration demand of SADm ~ 1.8 Nm3m-2h-1. The HRT of the MBBR
was 6 hours. Water quality parameters and treatment efficiencies are given in Table 1.
The pilot plants were equipped with National Instruments / LabVEIW data acquisition
units and online measurements using various sensors, i.e. temperature, pressure, flow
etc. All analyses were performed according to national or international standards. The
development of transmembrane pressure (TMP) was measured continuously using an
online pressure transducer connected to a National Instruments, FieldPoint (FP1000 and
FP-AI-110) unit, with the LabVIEW 6.1 and 8.2 data acquisition and analysis software.
TMP and temperature were logged for every two seconds. Data series where then
extracted from the raw data with a routine written in C++ software. Suspended solids
(SS) were analyzed by filtering through a Whatman GF/C 1.2 m glass microfiber filter
according to the Norwegian Standard NS 4733. Chemical Oxygen Demand (COD),
ammonia (NH4-N) and nitrogen (NO3-N) were measured with the Dr Lange LCK 114,
314, 303, 304 and 340 cuvette tests. For the Filtered Chemical Oxygen Demand
(FCOD) samples were first filtered with a Whatman GF/C 1.2 m filter. Particle size
distribution (PSD) analysis of the wastewater was done by using laser diffraction
spectroscopy (Beckman Coulter LS230).
97
Table 1. Quality parameters and performances of MBBR for low and high strength
wastewater
COD
[mg/L]
FCOD
[mg/L]
MLSS
[mg/L]
NH4-N
[mg/L]
NO3-N
[mg/L]
HRT [h]
Results
The impact of fouling rates observed within a membrane operating cycle as a function
of MLSS concentration in the membrane tank was measured. For the CM-MR
configuration, MLSS concentrations just under 400 mg/L up to around 3000 mg/L were
investigated. Results are presented in Figure 5, showing that an increase of MLSS
results in higher fouling rates, which is as expected based on reports from previous
studies of MBR processes. The correlation between MLSS concentration and fouling
rates is clear, indicating that a reduction in MLSS in the membrane reactor should result
in a more sustainable operation of the membrane filtration unit with considerably lower
fouling rates.
2,5E-04
2,0E-04
1,5E-04
y = 8E-08x - 2E-07
R2 = 0,9866
1,0E-04
5,0E-05
0,0E+00
0
500
1000
1500
2000
2500
3000
3500
MLSS [mg/L]
Figure 5. Example of fouling rate within an operating cycles as a function of MLSS for
CM-MR
98
Equation (6) was used to model and predict the MLSS concentrations for the alternative
reactor designs described above. For the CM-MR (Figure 4, a) ) the KS value was set to
one, while the steady state concentration for the SH-MR (Figure 4 b)) was fitted for
KS=3. During experiments with the SH-MR configuration it was observed that over time
the settled sludge in the sludge hopper had a tendency to float up and increase the
MLSS concentration around the membrane area. A modified reactor to handle the
floating sludge was therefore designed in order to separate floating sludge from
reaching the membrane area, MSH-MR in Figure 4 c). Measured steady state values of
MLSS in this configuration indicate that KS has a value of ~7,5. Experimental
verification of the models has been done and results are presented in Figure 6. A good
fit is apparent, confirming that it is possible to estimate the concentration of MLSS
inside the membrane reactor based on a given operating condition, quality of the MBBR
effluent into the membrane reactor and the membrane reactor geometry/design (i.e. KS
factor).
10000
MLSS (mg/L)
1000
100
MBBR effluent (Average value)
10
0
10
15
20
Time (hour)
25
30
35
40
Figure 6. Calculated and measured values for MLSS for CM-MR, SH-MR
and MSH-MR
The objective of this study was to determine the effect of the membrane reactor design
on controlling the MLSS concentration around the membrane area and thus reducing
membrane fouling rates. A comparison of the overall performance of the three reactor
designs is shown in Figure 7. The fouling rate, expressed as TMP development over
time, is shown both before and after backwashing during the start up period for all three
reactor designs. The differences between the two measurements (TMP) represent the
reversible fouling formed during a filtration cycle, while the observed TMP
development over time measured right after backwashing represents the irreversible
fouling. The reversible fouling rates within the production cycle after 32 hours of
operation were 6.18, 0.97, and 0.10 kPa/cycle for the CM-MR, SH-MR and MSH-MR,
respectively. The results clearly indicate a better performance and more sustainable
operation of the MSH-MR configuration due to enhanced solids control in the
membrane reactor.
99
0,30
0,25
0,20
0,15
SH-MR
0,10
MSH-MR
0,05
0,00
0
10
12
14
16
18
20
22
24
26
28
30
32
Time [hour]
Figure 7. TMP before and after backwash for CM-MR, SH-MR, MSH-MR.
Production flux 35LMH, and Backwash flux 38 LMH
The pilot plants were operated continuously over a period, setting a maximum TMP
level of 30 kPa (0.3 bar) as the condition to initiate chemically enhanced backwashing
(CEB) to remove the irreversible fouling. A comparison of the performance of the three
reactor configurations investigated is given in Figure 8, showing results for filtration of
high strength wastewater. The pilot plants were operated for a period of 7 days. The
CM-MR configuration was terminated after 3 days operation as it reached the set TMP
cut-off point within that period, while the other two configurations were still below this
point. The observed average fouling rates were 10 kPa/day, 3.57 kPa/day and 1.42
kPa/day for the CM-MR, SH-MR and MSH-MR respectively for the operating period
shown.
Table 2. Average values at steady state conditions measured around the membrane area
for three membrane reactors.
Low strength wastewater
High strength wastewater
CM-MR SH-MR MSH-MR CM-MR SH-MR MSH-MR
3400
1106
632
3630
1024
607
COD [mg/L]
306
325
105
448
198
105
131
91
78
225
119
108
FCOD [mg/L]
29
15
16
46
44
26
3110
757
460
2740
590
475
MLSS [mg/L]
316
234
66
541
99
136
HRT [h]
0.3
0.7
1.3
0.3
0.7
1.3
SRT [h]
7.5
20.8
33
7.5
20.8
33
Recovery [%]
95.5-96%
100
Results shown in Figure 8 confirm that the membrane reactor design plays an important
role in membrane fouling control. Even when most of the operating parameters (e.g.
HRT in MBBR, loading rates, intensity of membrane aeration, backwash/relaxation, net
flux, etc.) were varied during the experiments with low strength wastewater, the overall
membrane performance observed had the same trend regarding fouling dynamics. A
reduction of particulate matter around the membrane module i.e. reduced MLSS by
sedimentation in membrane reactor, can significantly improve membrane performance
i.e. reduced fouling.
0,4
0,35
CM-MR
0,3
TMP [bar]
0,25
0,2
SH-MR
0,15
0,1
MSH-MR
0,05
0
0
Time [day]
Figure 8. Overall TMP for CM-MR, SH-MR and MSH-MR at the constant flux of
22LMH
Lower concentrations of MLSS and COD (Table. 2) around the membranes as a
function of the modified rector designs results in a better membrane performance.
However, reduction in MLSS is not directly proportional to a reduction of fouling rates
(dTMP/dt). The characteristics of suspended matter around the membrane plays an
important role in membrane fouling, however, other foulants such as submicron
particles and SMPs have also been identified as having an impact on fouling
[10][11][13][15]. A reduction in these foulants by an enhanced membrane reactor
design is also shown to have a significant contribution to controlling and minimizing
fouling of the membrane. Soluble matter (FCOD) was reduced significantly in the
MSH-MR compared to the CM-MR (Table 2.).
However, the submicron particles and colloidal organic matter remain significant
foulants as reported in previous studies [1][2][17][18]. Previous studies have reported
the impact of reducing the colloidal submicron particles around the membrane and the
effect on membrane fouling. One strategy to achieve this is the integration of a
flocculation zone where submicron particles from the effluent of a MBBR are captured
by larger flocs that tend to settle [16]. Air-scouring for fouling control and mitigation in
submerged membrane reactors is also a challenge in that too high aeration intensities
101
may generate more colloidal material. A previous study demonstrated that there is a
tradeoff between increasing aeration intensities to prevent fouling and the formation of
submicron colloidal material caused by the high shear forces and particle breakage [17].
In this study the impact of introducing the MBBR effluent beneath the aeration device
in the membrane reactor to avoid floc breakage and induction of smaller, submicron
particles was re-investigated. Results show that it was possible to reduce the amount of
submicron particles around the membrane in the MSH-MR by designing the inlet point
beneath the membrane in a flocculation zone, and below the membrane aerator (Figure
9).
9
8
7
Number [%]
6
5
4
3
inlet
2
1
0
0.01
0.10
1.00
Size [m]
102
Conclusions
Modification of the membrane reactor design in a BF-MBR process is a potential tool to
improve the overall performance of the treatment process. The alternative designs
investigated in this study included introducing an integrated flocculation zone in the
membrane reactor coupled with a sedimentation zone beneath the membrane module.
The modified membrane rector design provided a significantly lower concentration of
MLSS and COD around the membranes, and subsequently a more sustainable
membrane performance due to much lower overall fouling rates. Reduction in MLSS is
not directly proportional to a reduction of fouling rates (i.e. dTMP/dt). The
characteristics of suspended matter around the membrane and other foulants also play
an important role in membrane fouling, in particular the submicron colloidal fraction.
This study has demonstrated that a reduction in these foulants by an enhanced
membrane reactor design is a significant contribution to controlling and minimizing
membrane fouling. A simple model has been proposed for calculating and predicting
steady-state values of MLSS inside the membrane reactor as a function of a given
membrane reactor design. This has been done by introducing a separation coefficient
(Ks) which is a function of reactor design, i.e. hydrodynamic conditions, integration of a
flocculation zone, sludge hopper etc. Further development and refinement of the model
by determining adequate expressions for KS will be investigated with the aim of
developing a design tool for improved membrane reactor designs for the BF-MBR
process.
Acknowledgements
Financial support by: Norwegian Research Council (NFR) and EUROMBRA research
project supported by the European Commission under the Sixth Framework
Programme, Contract No. 018480; Krger Kaldnes, Norway, for support with the
biofilm reactor and ZENON Environmental Inc., Canada, for supplying the membrane
modules.
References.
[1]
[2]
[3]
[4]
[5]
103
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
I.S. Chang and S.N. Kim, Wastewater treatment using membrane filtration
effect of biosolids concentration on cake resistance, Process Biochemistry, 40 (34) (2005) 1307-1314
S.S Han, T.H. Bae, G.G Jang. and T.M. Tak, Influence of sludge retention time on
membrane fouling and bioactivities in membrane bioreactor system, Process
Biochemistry, 40 (7) (2005) 2393-2400
C. Psoch and S. Schiewer, Anti-fouling application of air sparging and
backflushing for MBR, Journal of Membrane Science 283 (12) (2006) 273280.
R.S. Trussell, R.P. Merlo, S.W. Hermanowicz and D. Jenkins, Influence of mixed
liquor properties and aeration intensity on membrane fouling in a submerged
membrane bioreactor at high mixed liquor suspended solids concentrations, Water
Research 41 (5) (2007) 947958.
P. Le-Clech, V. Chen, and T.A.G. Fane, Fouling in membrane bioreactors used in
wastewater treatment, Journal of Membrane Science 284(1-2) (2006) 17-53.
F.G. Meng, S.R. Chae, A. Drews, M. Kraume, H.S. Shin and F.L. Yang, Recent
advances in membrane bioreactors (MBRs): membrane fouling and membrane
material, Water Research 43 (2009) 14891512.
L. Defrance, M. Y. Jaffrin, B. Gupta, P. Paullier, and V. Geaugey, Contributions
of various constituents of activated sludge to membrane bioreactor fouling.
Biores. Technol., 73 (2000) 105-112.
S. Rosenberger, C. Laabs, B. Lesjean, R. Gnirss, G. Amy; M. Jekel, and J.C
Schrotter, Impact of colloidal and soluble organic material on membrane
performance in membrane bioreactors for municipal wastewater treatment, Water
Res. 40 (2006) 710720.
Z. Ahmed, J. Chob, B.R. Limc, K.G. Songa and K.H. Ahna, Effects of sludge
retention time on membrane fouling and microbial community structure in a
membrane bioreactor, Journal of Membrane Science, (2007) 287(2) 211-218.
I. Ivanovic, T.O. Leiknes and H. degaard, Influence of loading rates on
production and characteristics of retentate from a biofilm membrane bioreactor
(BF-MBR). Desalination 199 (2006) 490-492.
I. Ivanovic, T.O. Leiknes and H. degaard, Fouling control by reduction of
submicron particles in a BF-MBR with an integrated flocculation zone in the
membrane reactor. Separation science and technology 43.(7) (2008) 1871-1883.
I. Ivanovic and T.O. Leiknes, Impact of aeration rates on particle colloidal
fraction in the biofilm membrane bioreactor (BF-MBR). Desalination 231 (2008)
182-190.
C. Sun, T.O. Leiknes, J. Weitzenbck and B. Thorstensen, The effect of bilge
water on a BiofilmMBR process in an integrated shipboard wastewater
treatment system, Desalination 236, (2009) 56-64.
104
105
122
123
124
Abstract
The effectiveness of five additives, two iron based and two alum based coagulants and
one modified cationic polymer was investigated in relation to reduction of fouling rates
in biofilm-MBR. Additionally, the amount of colloidal organic matter in terms of
soluble microbial products (SMP), dissolved organic carbon (DOC) and filtered
chemical oxygen demand (FCOD) removal, retained by membrane was related to
measured membrane fouling rates. Optimal dosage was defined based on maximum
values of coefficient additive utilization (CAU). Iron chloride at higher chosen dosage
showed that fouling could be reduced up to seven times, where for polymerized alum
chloride it was measured reduction of about three times. The iron chloride coagulant
performed better than iron chloride sulfate, in terms of fouling reduction. Higher
basicity of polymerized inorganic coagulant did not result in improved membrane
performance. Modified cationic polymer showed good potential in instantaneous
fouling reduction, however continuous dosing strategy was found difficult to use
without thorough monitoring of the system performance. Reduction in fouling rates
relates better to reduction of SMPcarbohydrates, than to SMPprotein and DOC. Also FCOD
was seen as good potential fouling predictor parameter. Synergetic effect of high total
phosphorus removal rates and reduced fouling rates give advantage of iron chloride
coagulant over the others tested additives in this study.
Key words. Biofilm MBR, fouling control, colloidal organic matter, coagulation
Introduction
The objective of this study is to investigate the effectiveness of different additives
commonly reported as filterability enhancers and fouling reducers in the MBR
technology on the overall performance in a biofilm-MBR (BF-MBR).
The BF-MBR is an alternative concept to conventional MBR, where a biofilm reactor is
employed instead of an activated sludge reactor. Several advantages of this approach
were previously reported; e.g. no need for biomass/sludge recirculation, significantly
lower concentration of MLSS and low viscosity of biofilm effluent giving lower energy
consumption for membrane aeration and less or no membrane module
sludging/clogging problems [1-3]. However, membrane fouling caused by suspended
and colloidal matter remains a major challenge in development of this concept, which is
also common for conventional MBR and other membrane systems [4-5].
125
126
127
128
From Figures 2-4 an optimum dosage not higher than 25 ppm for metal coagulants and
not higher than 10 ppm for cationic polymer was defined, if all four parameters are to be
taken into consideration with equal importance. Reduction of solids smaller then 1.2m
was followed by enlargement of suspended matter (>1.2m), was observed for all
applied dosages for both metal coagulants. This result indicates that colloidal matter, in
general, is additionally reduced by flocculation and adsorption by suspended matter (i.e.
fraction > 1.2m), suggesting this is a significant mechanism for improved
performance. Subsequently, the highest CAU [mg of submicron matter reduction/mg
additive] value was again used as the criteria for optimal dosage. Dosages of 15 ppm for
alum Figure 2b, and 10 ppm for iron Figure 3b were chosen as optimal and for the
polymer was roughly estimated as 10 ppm since no significant changes in upper micron
and submicron solids concentration were observed, Figure 4b.
Based on the results and analysis of the jar tests, to test possible improvements in
membrane performances experimental trials with the pilot plant were performed using
dosages of 9 and 22.5 ppm for both metal coagulants, iron chloride and alum chloride,
and 13.5 ppm for alternative iron chloride sulfate and alum chloride with high basicity.
For the cationic polymer trials were conducted with 45 ppm, and then gradually
decreasing the applied dose. Based on defined optimal dosages goal of pilot plant tests
was to compare which coagulant type (iron or alum based) performs better at lower and
higher dosages, (by comparing iron chloride and alum chloride), if there is a significant
difference in membrane performances between two iron type of inorganic coagulant (by
comparing iron chloride and iron chloride sulfate) and does higher basicity of
polymerized inorganic coagulant effects filterability (by comparing alum chloride and
alum chloride with high basicity). Also intention was to evaluate potential of modified
cationic polymer in BF-MBR and to additionally compare its efficiency to inorganic
coagulants. Reduction in fouling rates that refer to overall measured fouling was related
to applied dosages and reduction in amount of colloidal organic matter.
Table 1. Characteristics of inlet water, biofilm reactor effluent and removal rates
Inlet [mg/L]
MBBR effluent [mg/L]
MLSS
COD
176.48
614.50
(41.22) (72.47)
NH4-N
46.29
(8.21)
191.40
326.50
28.54
(39.65) (39.60) (15.81)
-8.45
46.87
FCOD
SMPp*
316.80
66.80
Inlet [mg/L]
(54.48) (10.59)
62.34
26.52
MBBR effluent [mg/L]
(7.35) (4.54)
80.32
60.31
R ***[%]
R [%]
38.34
SMPc*
10.25
(1.66)
8.57
(1.89)
16.42
TN
tot-P
52.04
22.30
(11.03) (3.21)
40.11
(7.93)
22.92
31.82
DOC
UV245**
91.78
0.69
(22.28) (0.15)
23.84
0.41
(3.30) (0.06)
74.03
-
* - refer to total amount of proteins and carbohydrates measured by colorimetric methods [21][22]
** - unit [cm-1]
***-R refer to removal rates
129
15.20
(2.67)
CAU
0.02
0.01
0
10
20
30
40
50
Dosage [ppm]
100
200
a)
100
0.08
0.07
>1.2m
0.06
0.05
0.2m<1.2m
50
0.04
CAU
Solids [%]
75
0.03
25
0.02
0.01
0
0
10
20
30
40
50
100
Dosage [ppm]
b)
Figure 2. Removal efficiency - jar test for alum: a) soluble organic compounds,
b) solids
130
0.03
FCOD
DOC
SMPp
SMPc
CAU
0.02
0.01
0
10
20
30
40
50
Dosage [ppm]
100
200
a)
100
0.08
0.07
0.06
0.05
>1.2m
50
0.04
0.2m<1.2m
CAU
Solids [%]
75
0.03
25
0.02
0.01
0
0
10
30
50
70
100
300
Dosage [ppm]
b)
Figure 3. Removal efficiency - jar test for iron: a) soluble organic compounds
b) solids
131
0.03
FCOD
DOC
SMPp
SMPc
CAU
0.02
0.01
0
10
30
50
70
100
Dosage [ppm]
300
500
a)
100
0.08
0.07
0.06
0.05
>1.2m
50
0.04
0.2m<1.2m
CAU
Solids [%]
75
0.03
25
0.02
0.01
0
0
10
20
30
40
50
100
Dosage [ppm]
b)
Figure 4. Removal efficiency- jar test for polymer: a) soluble organic compounds
b) solids
* >1m refer to solids fraction higher than 1 m and 0.2<1.2 m refer to solids fraction between 0.2 and
1.2 m
The iron based coagulant showed better performance than the alum at higher dosages
(i.e. 22.5 ppm), giving average fouling rates of 0.58 and 1.47 mbar/h, which was 7 and 3
times lower than in control reactor. This was related to better SMPc reduction of iron
based coagulant, however not proportionally to observed reduction, again indicating
complexity of membrane fouling in BF-MBR. Iron chloride sulfate at dosage of 13.5
ppm showed poorer performance in comparison to iron chloride at lower dosage (i.e. 9
ppm), with almost no improvement in membrane performance (i.e. fouling rate 3.92
mbar/h).
132
Alum chloride with higher basicity at dosage of 13.5 ppm resulted in almost the same
fouling rates (i.e. 1.67 mbar/h) as regular alum chloride at lower dosage (i.e. 9 ppm),
indicating that higher basicity not necessarily would improve filterability in BF-MBR.
Alum chloride with higher basicity effected better MLSS aggregation which was
measured as lower MLSS values around membrane for about 25% and in the same
manner higher in retentate stream, but better DOC removal expressed thought CAU,
due to this feature have not been seen as suggested in drinking water applications [18].
Results with cationic polymer indicated that only thoroughly controlled dosage of this
polymer can give improvement. Continuous dosage in the beginning of the cycle gave
improvement on performance since fouling rate was 1.71 mbar/h in first 24 hours,
however, later sharp increments in TMP development suggested that the membrane was
fouled by the polymer itself. This was observed even at much lower dosages, bellow 9
ppm. Results suggested that continuous dosing is not an option, since eventually
overdose of polymer occurs which results in membrane fouling by polymer. Therefore,
the effect of intermittent dosing on membrane performance was further investigated.
Both membranes were operated at a higher flux of 30 L.m-2.h-1 , with higher recovery of
96% for 12 hours, at which a dosage of 100 and 300 mg (per L of volume reactor) was
instantaneously applied (Figure 5). The sharp TMP rise was stabilized and after three
hours, when TMP started to rise a continuous dose of 2.7 ppm and 0.9 ppm was applied
for the next 20 hours. The strategy of 100 mg + 2.7 ppm reduced fouling rate by seven
times for tested period, giving an indication that application of cationic polymer could
be highly beneficial, however proper dosing strategy is crucial for successful use of this
additive [14-15]. However, this finding has to be tested for longer period of time and
further refinements of dosing strategies are required.
0.3
TMP [bar]
0.25
0.2
0.15
0.1
0.05
0
0
12
16
20
Operational time [hour]
24
28
32
Figure 5. Suggested strategy for polymer dosage for short term experiment
133
FCOD
SMPp
SMPc
DOC
i
(
)
i
(
SMPc: R2 = 0.7752
0.14
Fouling rate [bar/day]
0.12
0.1
)
SMPp: R2 = 0.6155
DOC: R2 = 0.5518
0.08
0.06
0.04
FCOD: R2 = 0.7217
0.02
0
0
10
20
30
40
50
Amount retained by membrane [mg/L]
60
134
135
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
136
137
138
Abstract
Improved performance of the BF-MBR process by addition of inorganic coagulants,
iron and alum based was studied with respect to reduction of surface area of colloidal
particles, enlargement of suspended flocs, and reducing the organic colloidal content.
Addition of coagulants in the membrane filtration stage of the BF-MBR had a positive
effect on improvement of mixed liquor filterability and dewaterability, resulting in
lower fouling rates. Modified jar tests were done in aerated beakers to determine
applicable dosages by investigating effects on flocculation index, number, size and
charge of colloidal particles, surface area of submicron particles, and solids and organic
colloidal content. Two dosages of 9 ppm and 22.5 ppm were chosen and tested in two
sets of experiments on two membrane rectors operated in parallel. Membrane fouling
was reduced in a similar manner for both lower dosages of alum and iron applied,
however, iron performed better at the higher dose. Alum was able to provide a higher
positive charge per unit of applied coagulant, and better addressed removal of colloidal
organic content measured as FCOD. Iron was more effective in enlarging surfaces of
flocs in the bulk phase, leading to more adsorption of submicron particles on floc
surfaces and probably promotion of a protective cake layer on the membrane surface.
Improvement in membrane performances was found to correlate to reduction of
measured surface area of particles in the size range 0.04 0.3m, enlargement of floc
surface area for particles > 8 m, as well as reduction of concentration of the colloidal
organic content. Total phosphorus removal was measured to be in the same range for
both coagulants added.
Keywords. Biofilm MBR, submicron particles, inorganic coagulant, colloidal fraction
Introduction
The BF-MBR is a concept similar to conventional MBR, where a bioreactor with
attached growth is employed instead of an activated sludge reactor [1]. Several
advantages of the purely biofilm based MBR approach over activated sludge MBR are;
1) zero energy demand for biomass recirculation, since biomass is attached to carriers
inside the bioreactor, 2) lower solids loads to membrane filtration unit, 3) lower aeration
for air scouring and easier cleaning due to lower viscosity and solids content, 4)
insignificant or no sludging problems of the membrane modules, 5) one more level of
freedom in designing the process due to the fact that operation of biological and
separation processes are decoupled [2][3]. However, membrane fouling is still the main
constraint for long term sustainable operation of a BF-MBR as is the case in activated
sludge based MBRs [4].
Coagulation and flocculation are commonly understood techniques that can assist
139
solid/liquid separation processes. In later years, coagulation has been used in MBR
technologies for improvement of filterability of the activated sludge and for fouling
control [4][5][6]. It has been demonstrated that different coagulants at optimum dosages
are able to reduce colloidal organic content measured as SMP (EPS), DOC or FCOD
[7][8], enlarge floc sizes [9][10] and reduce (bio)cake porosity [11][12]. Improvements
in membrane performances are reported as lower fouling rates, longer operational
cycles, higher operating fluxes or less intense mechanical and chemical cleaning
demand [9]. Polymerized inorganic coagulant has the ability to provide higher positive
charges than monomeric ones, and it was demonstrated that they are highly capable of
reducing membrane fouling [7][9]. Positive charge delivered by hydrolysis of the
coagulant in the bulk phase is partly consumed by colloids and partly adsorbed on the
surfaces of flocs. Therefore, charge neutralization, bridging and adsorption on floc
surfaces are the main mechanisms that are expected to induce reduction of submicron
(organic) matter [13].
Application of inorganic coagulants should be beneficial for overall performance
enhancement in BF-MBR, since it is expected that addition of a coagulant can reduce
the amount of submicron particles in solution, which is reported to be one of the main
foulants in this process [1][3]. Other positive effects are also to be expected such as
enlargement of floc sizes, improved filterability and dewaterability of the solids,
reduction of organic colloidal content and overall easier solids/liquid separation [14].
Negative effects of adding a coagulant are additional costs for the coagulant, higher
sludge production and higher treatment costs of chemically contaminated excess sludge
[15][16]. Potential toxicity of inorganic coagulants in activated sludge processes, in
particular iron based, [8][13] has no effect on the quantity and activity of biomass in a
biofilm process since the membrane separation stage is decoupled from the biological
treatment stage in a BF-MBR.
The aim of this study is to investigate the effect of two inorganic coagulants, iron and
polymerized alum based, on particle composition with respect to membrane fouling and
overall system performance. Employing additional equipment (i.e. pipe flocculator,
mixing chambers, flash and static mixers etc.) and additional energy for that purpose
was intentionally avoided by applying the coagulant directly to the membrane filtration
unit (Figure 1.).
Materials and methods
Two commercial metal based salt coagulants were chosen; iron based and polymerized
alum based. Jar tests were done for five different concentrations, chosen based on
reported values found in the literature [7][8][9] and preliminary jar tests with a shaken
flask. Prepared additives were dosed to mixed liquor taken from the membrane reactor
in concentrations of 10, 20, 30, 40 and 50 ppm metal. Additionally, two higher dosages
were applied in order to estimate iso-electric points, i.e. charge neutrality of colloids. A
modified jar test procedure was conducted in a 800 ml beaker (dxh=90x120mm)
aerated from the bottom with a 7.5 L/min coarse bubble aeration in order to provide
mixing conditions similar to that in the membrane reactor. Monitoring of the flocs
growth and stability was done by a Photometric Dispersion Analyzer (PDA 2000, Rank
140
Bros.Ltd., Cambridge, UK) with transparent tube of 3mm internal diameter and a flow
rate of 20 ml/min according to [17]. For each dosage, changes in floc structure were
measured and flocculation index logged over 30 minutes. MLSS and particle size
distribution (0.04 m-2000 m) was measured for each sample. The sample were
furthermore centrifuged @ 20000 rpm for 3.5 min and supernatant was taken for
measurement of colloidal organic content (i.e. FCOD), unimodal size distribution of
nano-particles (in range 3 nm to 1000 nm) and zeta potential. Total time for reaction for
each applied coagulant dose with water was one hour.
A schematic of the BF-MBR pilot plant used in this study is shown in Figure 1 and the
bioreactor configuration is described in previous studies [1][3]. Membranes were
operated continuously at constant flux of 25 L m-2 h-1 (without backwash or relaxation)
and with constant aeration of SADm ~ 1.8 Nm 3 m -2 h -1 . Recovery was set at 90 % and
HRT and SRT were 1 and 10.1 hour, respectively. Membranes were operated until TMP
increased up to 0.3 bar or 7 operational days, after which the membranes were
chemically cleaned. Two sets of experiments, Exp 1. and Exp 2., with lower and higher
chosen dosages by alternating type of coagulant. Coagulants were diluted with tap
water, and then continuously dosed using a computer controlled peristaltic pump (L/S
Easy-load II, Masterflex). The top of the membrane reactor was chosen as a coagulant
adding point since very good mixing and dispersion of applied additive was secured by
strong aeration for membrane air scouring purpose.
141
142
5
Al 10
4
F index
Al 30
3
Al 50
2
1
0
0
10
15
20
25
Time [min]
a)
5
F index
Fe 10
Fe 30
Fe 50
2
1
0
0
10
15
20
25
Time [min]
b)
Figure 2. Jar test results Flocculation index for: a) alum , b) iron
Taking into account the reduction of surface area of submicron particles ( <1m), the
optimal dose appears to be not lower than 20 ppm for alum and between 10-20 ppm for
iron, since no significant reduction of particle surface area was measured for higher
applied dosages. For the larger particles greater than one micron a similar conclusion
can be drawn (i.e. enlargement of flocs surface) (Figure 3a) and 3b).
143
3
Al 0
Al 10
Al 20
Al 30
Al 40
Al 50
Al 100
2,5
2
1,5
1
0,5
0
0,01
0,1
10
100
1000
Size [m]
a)
3
Fe 0
Fe 10
Fe 20
Fe 30
Fe 40
Fe 50
2,5
2
1,5
1
0,5
0
0,01
0,1
Size [m]
10
100
1000
b)
Figure 3. Jar test results Changes in distribution of particles surface area % :
a) alum , b) iron
Changes and growth in size of the nano particles in the supernatant, and consequently
their reduction in number, as a function of applied dosage is presented in Figure 4a) and
4b). A unimodal size at highest laser diffraction intensity had a peak at 260 (3.95) nm
for 40 ppm of alum (Figure 4a) ). For iron, the largest sizes were measured at 236.3
(5.51) nm for a dosage of 30 ppm (Figure 4 b)). The highest reduction in number of
counts (i.e. particles) per applied dosage was for a dosage of 10 ppm for both
coagulants. Alum performed slightly better at higher applied dosages compared to iron
(i.e. had higher reduction in number of counts).
144
260
Number of counts
Size [nm]
Number of counts
2E+14
250
1,5E+14
240
1E+14
230
5E+13
220
2,5E+14
210
Al 0
Al 10
Al 20
Al 30
Al 40
Al 50
Dosage
a)
260
Number of counts
Size [nm]
Number of counts
2E+14
250
1,5E+14
240
1E+14
230
5E+13
220
2,5E+14
210
Fe 0
Fe 10
Fe 20
Fe 30
Fe 40
Fe 50
Dosage
b)
Figure 4. Jar test results Number of particles and size at max intensity :
a) alum , b) iron
The stability of the colloids was determined by measuring zeta potentials as a function
of applied dosages. No significant destabilization of the colloids was observed for most
applied dosages. When additional coagulant was applied iso-electrical charge inversion
took place and iso-electric points where estimated by linear interpolation for dosages of
92 ppm alum and 155 ppm iron (Figure 5). These results suggest that charge
neutralization is probably not a key destabilizing mechanism and therefore not be
appropriate criteria for estimating applicable doses. Adsorption and/or bridging thus
appears to be the dominant mechanisms for reduction of the colloidal fraction in the
complex suspension of flocs, large unassociated particles, colloids and solutes as found
in the bulk phase in BF-MBR processes [9].
Sludge enlargement measured as MLSS and reduction of organic colloidal content (i.e.
FCOD) for applied dosages are shown in Figure 6. Alum was found to be more
145
effective in reducing the amount of organic colloidal matter with increasing dose
compared to iron, while both coagulants increased the sludge content to some extend
proportionally to applied dosages.
Pilot plant results
Studies on a BF-MBR pilot plant were further conducted with two sets of experiment,
identified as Exp 1. and 2. Two different doses of 9 ppm and 22.5 ppm, for both
coagulants tested, where chosen based on jar test results in order evaluate the
effectiveness with respect to changes in particle composition and overall membrane
filtration performance. In Exp 1. one reactor was operated with 9 ppm alum and the
other with 22.5 ppm of iron. After that membranes were chemically cleaned and Exp 2.
was then conducted with 9 ppm of iron and 22.5 ppm of alum.
0
0
50
100
150
200
-5
Alum
-10
Iron
-15
Dosage [mg/L]
20
1000
900
800
700
600
500
400
300
200
100
0
30
40
50
60
250
200
150
FCOD
Al MLSS
100
Fe MLSS
Al FCOD
50
Fe FCOD
0
0
10
20
30
40
50
Dosage [ppm]
146
FCOD [mg/L]
MLSS [mg/L]
Prior to these tests the pilot plant was operated for a two week period without coagulant,
both to synchronize the performances of the membrane reactors and to minimize
uncertainties in operation, and to obtain reference TMP values for operation without
additives.
The pilot plant was fed with municipal wastewater from a combined sewer system in
the municipality of Trondheim, Norway. Inlet water characteristics are shown in Table
1. The bioreactors operated at HRT of 4 hours with a stable performance with respect to
COD, FCOD, and tot-P removal rates (Table 1). Water temperature was close to room
temperature, on average 18.5 21.5 oC, during the whole period.
Results from the study with the additives are summarized in Table 2. Alum was found
to have a better ability to bind colloidal organic matter compared to iron for both chosen
dosages (Table 2). This is in agreement with similar results are reported in other studies
[9]. Removal of colloidal organic matter by membrane filtration with addition of
coagulant (highest to lowest) was in the order Al 22.5 ppm; Al 9 ppm; Fe 22.5 ppm; Fe
9 ppm.
Table 1. Average values for Exp 1. and Exp 2. of inlet water and biofilm reactor
effluent
Inlet
[mg/L]
BF eff.
[mg/L]
Rem. rates
[%]
MLSS
[mg]
176.48
(41.22)
191.40
(39.65)
COD
[mg/L]
614.50
(72.47)
308.50
(39.60)
FCOD
[mg]
316.80
(54.48)
62.34
(7.35)
TP
[mg]
22.30
(3.21)
15.20
(2.67)
-8.45
46.87
80.32
31.82
147
Results of measured surface areas of particulate content in the range from 0.04 to 2000
m for four sampling points are shown in Figures 7 a) and 7b). The biological treatment
significantly changes the characteristics of the particulate matter in the wastewater,
particularly with a large reduction of particles in the size range 0.3 to 7-8 m.
Additionally, a clear increase in percent surface area of particles in the range 0.04 to 0.3
and flocs > 8 m was measured.
Table 2. Average values for Exp 1 and Exp 2 of bulk phase, permeate and concentrate
Bulk phase (around the membrane)
Exp 1
Exp 2
BF
eff.
Fe
9
Al
22.5
BF
eff.
Al
9
Fe
22.5
MLSS
[mg/L]
184.00
28.86
1040.00
60.83
1271.67
391.16
198.00
54.11
1315.56
140.33
2192.22
236.46
COD
[mg/L]
307.67
12.42
1206.67
88.12
1205.67
362.4
315.33
50.81
1503.33
82.03
2409.00
326.51
Permeate
FCOD
TP
[mg/L] [mg/L]
Exp 1
Exp 2
BF
eff.
Fe
9
Al
22.5
BF
eff.
Al
9
Fe
22.5
FCOD
[mg/L]
59.30
12.21
71.20
18.14
43.00
14.39
65.30
7.45
63.43
6,26
58.67
3.85
CSTn
[s/g/L]
68.13
40.56
54.82
41.26
Concentrate
MLSS
[mg/L]
COD
[mg/L]
26.20
9.75
27.17
4.95
4.88
0.15
1.46
0.64
2186.1
889.5
3308.9
280.4
2674.0
520.1
3068.0
429.0
32.23
6.17
29.00
4.51
4.04
0.41
1.66
0.30
4046.0
1396.5
3469.7
753.7
4313.3
1926.4
3165.00
570.3
The highest surface area % was found for particles in the size range 0.1 to 0.3 m for
Exp 1., and 0.08 to 0.3 for Exp 2, with respective peak values of 1.75 and 2.6 m. These
148
results show that the biofilm effluent during Exp 1. had more favorable characteristics
with respect to lower submicron colloidal content than in Exp 2. However, added
coagulants where not able to significantly change the fingerprint of the BF effluent.
Both coagulants applied were capable of reducing the surface area of the submicron
content by increasing the number of larger particles and thus increasing the floc surface
area of particles in the upper micron range. Effectiveness of the added coagulant with
respect to shifting towards a more favorable distribution in the submicron range was in
the order Fe 22.5 ppm; Al 22.5 ppm ~ Al 9 ppm; Fe 9 ppm. Enlargement in flocs
surfaces due to presences of a coagulant followed the same order.
4
Inlet
BF eff.
Al 9ppm
Fe 22,5ppm
3,5
3
2,5
2
1,5
1
0,5
0
0,01
0,1
10
100
1000
Size [m]
a)
4
Inlet
BF eff.
Fe 9ppm
Al 22,5 ppm
3,5
3
2,5
2
1,5
1
0,5
0
0,01
0,1
10
100
1000
Size [m]
b)
Figure 7. Surface area distribution for four sampling points for a) Exp 1. and b) Exp 2.
Both alum and iron at the lower dosage (i.e. 9 ppm), performed similarly giving average
fouling rates of ~1.5 and ~1.7 mbar/h, respectively. This was lower than average fouling
rates observed during the control runs, measured at ~ 4 mbar/h. The iron based
coagulant showed better performance than the alum at the higher dosages (i.e. 22.5
149
ppm), giving average fouling rates of 0.6 and 1.5 mbar/h (Figure 8a) and 8b)). The
ability of the coagulants to reduce fouling based on type and dose was found to be in the
order Fe 22.5 ppm; Al 22.5 ppm; Al 9ppm ~ Fe 9ppm. These results suggest that even
though higher solids concentration was measured (Table 2), these probably consisted of
lighter weight flocs with high surfaces area (in range 3 80 m (Figure 7a)), leading to
a higher adsorption and reduction of the submicron particles.
y = 3,97x + 36,83
R2 = 0,99
300
y = 1,50x + 52,91
R2 = 0,97
TMP [mbar]
250
y = 0,58x + 49,30
R2 = 0,97
200
150
Ref
100
Al 9ppm
50
Fe 22.5ppm
Li
0
0
20
40
60
80
100
120
140
(R f)
160
180
a)
300
y = 3,97x + 36,83
R2 = 0,99
y = 1,69x + 18,33
R2 = 0,94
TMP [mbar]
250
y = 1,47x + 29,75
R2 = 0,97
200
150
Ref
100
Al 22.5ppm
50
Fe 9ppm
Li
0
0
20
40
60
80
100
120
140
(R f)
160
180
b)
Figure 8. The TMP development for a) Exp 1. and b) Exp 2
Additionally, it is highly probable that the suspension with large surface area and light
flocs has formed a protective cake layer on the membrane surface, reducing the fouling
potential of the colloidal (organic) content in bulk phase and forming a dynamic cake
layer with less hydraulic resistance. It is also important to note that the BF effluent
during the Exp 1. had more favorable characteristics (i.e. less submicron particles) than
during Exp. 2, and firm conclusions can therefore not be made based on this set of
150
experiments. Further studies on the nature and characteristics of the cake layer formed
on the membrane are required.
Conclusions
Two inorganic coagulants were applied in order to investigate the potential of reducing
membrane fouling in a BF-MBR process by coagulation. Extensive trials using a
modified jar test suggested that applicable dosages should be in the range of 10 to 25
ppm for pilot plant trials, based on FI, reduction in particle number, charge, size and
reduction in surface area of submicron particles. Pilot studies with addition of both
coagulants at the lower dosage range resulted in similar membrane performances
improvement (i.e. less fouling). The best improvement in membrane performance was
observed for the higher dosage range using iron chloride, while polymerized alum gave
a lower improvement in performance at the same applied dosage. Results suggest that
the response observed was related to a higher reduction in particle surface area of the
submicron particles in the size range 0.04 0.3m, and the lower organic colloidal
content.
The improved membrane performances observed at the higher iron dosage as assumed
to be related to the formation of lighter flocs with large surface areas which can form a
protective cake layer on the membrane surface thereby reducing fouling by the
submicron fractions. Both coagulants were able to satisfactory remove phosphorus.
Based on the testes conducted in this study the iron chloride coagulant at doses around
20 ppm appears to be the coagulant of choice for a BF-MBR process reactor. Further
studies are required to optimize the use of coagulant and to get a better understanding of
the particle properties and characteristics best suited to give an overall improvement on
membrane filtration performance in a BF-MBR process.
Acknowledgements
Financial support by: the Norwegian Research Council (NFR), Krger Kaldnes,
Norway, for support with the biofilm reactor, Zenon Environmental Inc., Canada, for
supplying the membrane modules, Kemira, Finland for support with chemicals.
References.
[1] T.O. Leiknes and H. degaard, The development of a biofilm membrane
bioreactor , Desalination, 202 (2007) 135-143.
[2] J. Phattaranawik and T.O. Leiknes, Double-deck aerated biofilm membrane
bioreactor with sludge control for municipal wastewater treatment, AIChE Journal,
55 (2009) 1291-1297.
[3] I. Ivanovic and T.O. Leiknes, Membrane reactor design as a tool for better
membrane performance in a biofilm MBR (BF-MBR). Desalination and water
treatment: (2010) in press
[4] A. Drews, Membrane fouling in membrane bioreactors- Characterization,
contradictions, cause and cures. Journal of Membrane Science 363(1-2): (2010). 128.
151
152
153
154
155
Abstract
Submicron particles represent one of the major foulants in the biofilm membrane reactor
BF-MBR. Reduction of the amount of submicron particles (colloids) adjacent to the
membrane is one measure in order to provide better fouling control in BF-MBR
systems. A submerged hollow fiber (Zenon Zeeweed) membrane reactor was redesigned
by introducing a flocculation zone below the aeration device of the membrane module.
This resulted in reduction of submicron particles around the membrane from 8.2% to
6.9 %, expressed in differential number percentage. The size of the most abundant
particle fraction consequently increased from 0.70 to 0.84 m. Furthermore, the
modified membrane reactor design provided longer operational cycles, >40% reduction
of suspended solids around the membrane, and improved retentate/concentrate
characteristics, i.e. dewaterability (CST), settleability (SVI/SSV) and filterability (TTF).
156
Introduction
A hybrid biofilm membrane reactor based on the moving-bed-biofilm reactor (MBBR)
combined with solids separation in an immersed membrane reactor (IMR) has been
investigated for the treatment of municipal wastewater. This process enables the design
of compact treatment plants with high biodegradation efficiencies that produce high
quality effluents. Retentate characteristics play an important role in the efficiency of the
solids separation and subsequently the overall performance of the membrane unit. The
aim of this study was to investigate and compare characteristics of the retentate in the
membrane reactor as a function of organic loading rates in the bioreactor of this BFMBR process.
157
158
159
160
161
Poster presentations
Leiknes, TorOve; Ivanovic, Igor; degaard, Hallvard.
162